)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     wang.weiya \u003cwang.weiya@zte.com.cn\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-01-20 10:49:34 +0800"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"blueprint alarm-deduction"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Change-Id: I0126a3b19ee37b2b80315c799cb6b7f62556c7b9"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"5a3905b3_db38c8d5","line":7,"range":{"start_line":7,"start_character":0,"end_line":7,"end_character":25},"updated":"2017-01-24 06:37:30.000000000","message":"Please change it to:\nImplements: blueprint alarm-deduction\n\nThis way, a link to the gerrit topic will automatically be added in the launchpad blueprint","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"9a6a567500fb33e042535a5cc406624e41d18ce4","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Commit:     wang.weiya \u003cwang.weiya@zte.com.cn\u003e"},{"line_number":5,"context_line":"CommitDate: 2017-01-20 10:49:34 +0800"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"blueprint alarm-deduction"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Change-Id: I0126a3b19ee37b2b80315c799cb6b7f62556c7b9"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"5a3905b3_e993bafe","line":7,"range":{"start_line":7,"start_character":0,"end_line":7,"end_character":25},"in_reply_to":"5a3905b3_db38c8d5","updated":"2017-01-25 07:50:01.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"}],"specs/pike/alarm-deduction.rst":[{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":4,"context_line":"\r"},{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode\r"},{"line_number":6,"context_line":"\r"},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":8,"context_line":"alarm deduction\r"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":10,"context_line":"\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_20019906","line":7,"range":{"start_line":7,"start_character":15,"end_line":7,"end_character":29},"updated":"2017-01-24 06:37:30.000000000","message":"The line is too long","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ace79cc2187ff81a32defa0e26e35cd0e33629b6","unresolved":false,"context_lines":[{"line_number":4,"context_line":"\r"},{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode\r"},{"line_number":6,"context_line":"\r"},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":8,"context_line":"alarm deduction\r"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":10,"context_line":"\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a430d35_1ff933c9","line":7,"range":{"start_line":7,"start_character":15,"end_line":7,"end_character":29},"in_reply_to":"3a461143_ae59905c","updated":"2017-02-03 01:03:27.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":4,"context_line":"\r"},{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode\r"},{"line_number":6,"context_line":"\r"},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":8,"context_line":"alarm deduction\r"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":10,"context_line":"\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_75c6ca4a","line":7,"range":{"start_line":7,"start_character":15,"end_line":7,"end_character":29},"in_reply_to":"5a3905b3_20019906","updated":"2017-01-25 06:55:30.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"329b5fd13cb4b030f2901f92e65efb2164dadd72","unresolved":false,"context_lines":[{"line_number":4,"context_line":"\r"},{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode\r"},{"line_number":6,"context_line":"\r"},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":8,"context_line":"alarm deduction\r"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\r"},{"line_number":10,"context_line":"\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_ae59905c","line":7,"range":{"start_line":7,"start_character":15,"end_line":7,"end_character":29},"in_reply_to":"5a3905b3_75c6ca4a","updated":"2017-01-26 18:33:16.000000000","message":"Now one \u003d is missing :-) :-) \nSorry to mention it again... but it will affect the way the blueprint will look under http://specs.openstack.org/openstack/vitrage-specs/.\nI suggest that you use rst.ninjs.org to verify the format.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":10,"context_line":"\r"},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction\r"},{"line_number":12,"context_line":"\r"},{"line_number":13,"context_line":"Vitrage should support alarm deduction.\r"},{"line_number":14,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed).\r"},{"line_number":15,"context_line":"\r"},{"line_number":16,"context_line":"\r"},{"line_number":17,"context_line":"Problem description\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_1b8b3069","line":14,"range":{"start_line":13,"start_character":0,"end_line":14,"end_character":104},"updated":"2017-01-24 06:37:30.000000000","message":"Vitrage already supports alarm deduction (e.g. in case of a host alarm, vitrage creates deduced alarms on the instances).\n\nWhat you would like to support is a special case of alarm deduction - that Vitrage will raise an alarm that should be raised by an external monitor, before the external monitor raises it. And then you would like to merge/replace vitrage alarm.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ace79cc2187ff81a32defa0e26e35cd0e33629b6","unresolved":false,"context_lines":[{"line_number":10,"context_line":"\r"},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction\r"},{"line_number":12,"context_line":"\r"},{"line_number":13,"context_line":"Vitrage should support alarm deduction.\r"},{"line_number":14,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed).\r"},{"line_number":15,"context_line":"\r"},{"line_number":16,"context_line":"\r"},{"line_number":17,"context_line":"Problem description\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a430d35_3ff6efd7","line":14,"range":{"start_line":13,"start_character":0,"end_line":14,"end_character":104},"in_reply_to":"3a461143_ae30b002","updated":"2017-02-03 01:03:27.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":10,"context_line":"\r"},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction\r"},{"line_number":12,"context_line":"\r"},{"line_number":13,"context_line":"Vitrage should support alarm deduction.\r"},{"line_number":14,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed).\r"},{"line_number":15,"context_line":"\r"},{"line_number":16,"context_line":"\r"},{"line_number":17,"context_line":"Problem description\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_f9282b88","line":14,"range":{"start_line":13,"start_character":0,"end_line":14,"end_character":104},"in_reply_to":"5a3905b3_1b8b3069","updated":"2017-01-25 06:55:30.000000000","message":"Yes, this is exactly what I mean.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"329b5fd13cb4b030f2901f92e65efb2164dadd72","unresolved":false,"context_lines":[{"line_number":10,"context_line":"\r"},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction\r"},{"line_number":12,"context_line":"\r"},{"line_number":13,"context_line":"Vitrage should support alarm deduction.\r"},{"line_number":14,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed).\r"},{"line_number":15,"context_line":"\r"},{"line_number":16,"context_line":"\r"},{"line_number":17,"context_line":"Problem description\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_ae30b002","line":14,"range":{"start_line":13,"start_character":0,"end_line":14,"end_character":104},"in_reply_to":"5a3905b3_f9282b88","updated":"2017-01-26 18:33:16.000000000","message":"When you say \"Vitrage should support alarm deduction\", it sounds to me that currently Vitrage does not support it. Which is not the case","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":25,"context_line":"\r"},{"line_number":26,"context_line":"We could identify the raised alarm B by vitrage and the reported alarm B by other datasource as the same alarm. In other words, we could merge the two alarms. We could just replace the vitrage raised alarm to datasource reported alarm.\r"},{"line_number":27,"context_line":"\r"},{"line_number":28,"context_line":"1. Templates should specify the datasource type when the action is \"raise-alarm\"(Default type is vitrage).\r"},{"line_number":29,"context_line":"2. Alarms from evaluator_queue should be added the ``is_deduced`` property, and a new transformer may be needed.\r"},{"line_number":30,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph. If yes just repalce the vertex. Otherwise, continue the original process.\r"},{"line_number":31,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too. If alarm B recovered message is received later, just discard the message.\r"},{"line_number":32,"context_line":"\r"},{"line_number":33,"context_line":"\r"},{"line_number":34,"context_line":"Alternatives\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_7b40b407","line":31,"range":{"start_line":28,"start_character":0,"end_line":31,"end_character":146},"updated":"2017-01-24 06:37:30.000000000","message":"I\u0027m not sure this will work. I think it\u0027s best if we discuss it in a design session. I can add more detailed comments later this week.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":25,"context_line":"\r"},{"line_number":26,"context_line":"We could identify the raised alarm B by vitrage and the reported alarm B by other datasource as the same alarm. In other words, we could merge the two alarms. We could just replace the vitrage raised alarm to datasource reported alarm.\r"},{"line_number":27,"context_line":"\r"},{"line_number":28,"context_line":"1. Templates should specify the datasource type when the action is \"raise-alarm\"(Default type is vitrage).\r"},{"line_number":29,"context_line":"2. Alarms from evaluator_queue should be added the ``is_deduced`` property, and a new transformer may be needed.\r"},{"line_number":30,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph. If yes just repalce the vertex. Otherwise, continue the original process.\r"},{"line_number":31,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too. If alarm B recovered message is received later, just discard the message.\r"},{"line_number":32,"context_line":"\r"},{"line_number":33,"context_line":"\r"},{"line_number":34,"context_line":"Alternatives\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_59b69ff9","line":31,"range":{"start_line":28,"start_character":0,"end_line":31,"end_character":146},"in_reply_to":"5a3905b3_7b40b407","updated":"2017-01-25 06:55:30.000000000","message":"ok, we will discuss it later.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Alternatives\r"},{"line_number":35,"context_line":"------------\r"},{"line_number":36,"context_line":"\r"},{"line_number":37,"context_line":"None.\r"},{"line_number":38,"context_line":"\r"},{"line_number":39,"context_line":"\r"},{"line_number":40,"context_line":"Data model impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_3b08ecd9","line":37,"range":{"start_line":37,"start_character":0,"end_line":37,"end_character":5},"updated":"2017-01-24 06:37:30.000000000","message":"There is an alternative: have both vitrage alarm and the monitor alarm, and connect them with a \u0027equivalent\u0027 edge. We should discuss the advantages and disadvantages of each solution.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ace79cc2187ff81a32defa0e26e35cd0e33629b6","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Alternatives\r"},{"line_number":35,"context_line":"------------\r"},{"line_number":36,"context_line":"\r"},{"line_number":37,"context_line":"None.\r"},{"line_number":38,"context_line":"\r"},{"line_number":39,"context_line":"\r"},{"line_number":40,"context_line":"Data model impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a430d35_7f0077ab","line":37,"range":{"start_line":37,"start_character":0,"end_line":37,"end_character":5},"in_reply_to":"3a461143_c1194d7b","updated":"2017-02-03 01:03:27.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Alternatives\r"},{"line_number":35,"context_line":"------------\r"},{"line_number":36,"context_line":"\r"},{"line_number":37,"context_line":"None.\r"},{"line_number":38,"context_line":"\r"},{"line_number":39,"context_line":"\r"},{"line_number":40,"context_line":"Data model impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_d5021e72","line":37,"range":{"start_line":37,"start_character":0,"end_line":37,"end_character":5},"in_reply_to":"5a3905b3_3b08ecd9","updated":"2017-01-25 06:55:30.000000000","message":"Well, It sounds a good alternative, and all information are contained in the graph without any missing. But it may be more complex:\n1. The templates would be more complex, because vitrage alarm and the monitor alarm are two different alarms in the graph. \n\nFor example, alarm A causes alarm B, and alarm B causes alarm C, we have to specify these relations:\n\ta)  A causes B\n\tb)  A causes B-Vitrage \n\tc)  B causes C\n\td)  B causes C-Vitrage\n\te)  B-Vitrage causes C-Vitrage\n\t\n2. It would be hard to connect vitrage alarm and the monitor alarm with a \u0027equivalent\u0027 edge. \n \nCurrently, monitor alarms are identified as \u0027ALARM:nagios:node-1:alarm-A\u0027, so we have to make sure that alarm name must be unified in vitrage, not only in there own monitor.\n    \n3. It may have more impact on performance because there are more vertex, and edges.","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"329b5fd13cb4b030f2901f92e65efb2164dadd72","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Alternatives\r"},{"line_number":35,"context_line":"------------\r"},{"line_number":36,"context_line":"\r"},{"line_number":37,"context_line":"None.\r"},{"line_number":38,"context_line":"\r"},{"line_number":39,"context_line":"\r"},{"line_number":40,"context_line":"Data model impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3a461143_c1194d7b","line":37,"range":{"start_line":37,"start_character":0,"end_line":37,"end_character":5},"in_reply_to":"5a3905b3_d5021e72","updated":"2017-01-26 18:33:16.000000000","message":"I agree to everything you said. But since the proposed change is also non-trivial, I think it makes sense to discuss also the alternative. I suggest that you write in the rst everything you wrote here, mention all the problems, and explain why you think it is not a good solution. This can be a good basis for our discussion later on","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":52,"context_line":"Versioning impact\r"},{"line_number":53,"context_line":"-----------------\r"},{"line_number":54,"context_line":"\r"},{"line_number":55,"context_line":"None\r"},{"line_number":56,"context_line":"\r"},{"line_number":57,"context_line":"\r"},{"line_number":58,"context_line":"Other end user impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_1bfed0ad","line":55,"range":{"start_line":55,"start_character":0,"end_line":55,"end_character":4},"updated":"2017-01-24 06:37:30.000000000","message":"There might be an impact if you change the template format","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":52,"context_line":"Versioning impact\r"},{"line_number":53,"context_line":"-----------------\r"},{"line_number":54,"context_line":"\r"},{"line_number":55,"context_line":"None\r"},{"line_number":56,"context_line":"\r"},{"line_number":57,"context_line":"\r"},{"line_number":58,"context_line":"Other end user impact\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_50c39c8a","line":55,"range":{"start_line":55,"start_character":0,"end_line":55,"end_character":4},"in_reply_to":"5a3905b3_1bfed0ad","updated":"2017-01-25 06:55:30.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"00e8b41efa1880ad9bba80e9ad28cb2bd2c07e83","unresolved":false,"context_lines":[{"line_number":96,"context_line":"Work Items\r"},{"line_number":97,"context_line":"----------\r"},{"line_number":98,"context_line":"\r"},{"line_number":99,"context_line":"None\r"},{"line_number":100,"context_line":"\r"},{"line_number":101,"context_line":"\r"},{"line_number":102,"context_line":"Dependencies\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_1bc33058","line":99,"range":{"start_line":99,"start_character":0,"end_line":99,"end_character":4},"updated":"2017-01-24 06:37:30.000000000","message":"There will be a few work items, so it\u0027s TBD","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"c9cee0cc9f79c47531ae90bfc6a66f142ca828bc","unresolved":false,"context_lines":[{"line_number":96,"context_line":"Work Items\r"},{"line_number":97,"context_line":"----------\r"},{"line_number":98,"context_line":"\r"},{"line_number":99,"context_line":"None\r"},{"line_number":100,"context_line":"\r"},{"line_number":101,"context_line":"\r"},{"line_number":102,"context_line":"Dependencies\r"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a3905b3_f5e0face","line":99,"range":{"start_line":99,"start_character":0,"end_line":99,"end_character":4},"in_reply_to":"5a3905b3_1bc33058","updated":"2017-01-25 06:55:30.000000000","message":"Done","commit_id":"96cae6850695fcd9462d86d5bff588a1397d917e"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d2268c10c72ee08322519690d62fa81ee9729688","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the deduced alarm must be a vitrage type alarm."},{"line_number":14,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will raise an alarm that should be raised by an external monitor, before the external monitor raises it."},{"line_number":15,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed)."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Problem description"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_14771f3b","line":15,"updated":"2017-02-05 02:51:20.000000000","message":"IMO, this line is enough for summary. We may move L13 and L14 to Problem Description","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ee482b27a4b4caab1b1dceb9d4978dbddb2408cb","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the deduced alarm must be a vitrage type alarm."},{"line_number":14,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will raise an alarm that should be raised by an external monitor, before the external monitor raises it."},{"line_number":15,"context_line":"This feature will allow us to get the deduced alarm that are not reported as expected (lost or delayed)."},{"line_number":16,"context_line":""},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Problem description"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_56decf3f","line":15,"in_reply_to":"1a430d35_14771f3b","updated":"2017-02-07 01:47:00.000000000","message":"Done","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d2268c10c72ee08322519690d62fa81ee9729688","unresolved":false,"context_lines":[{"line_number":18,"context_line":"Problem description"},{"line_number":19,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"There is a known fault model: alarm A causes an alarm B. When alarm A is reported, we hope we can deduced the alarm B at once, even if alarm B is delayed or could not be reported due to some reason(network problem, physical error and etc)."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_7494e336","line":21,"updated":"2017-02-05 02:51:20.000000000","message":"Wrap long lines manually at Column 79 so it will be easier to review in gerrit.","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ee482b27a4b4caab1b1dceb9d4978dbddb2408cb","unresolved":false,"context_lines":[{"line_number":18,"context_line":"Problem description"},{"line_number":19,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"There is a known fault model: alarm A causes an alarm B. When alarm A is reported, we hope we can deduced the alarm B at once, even if alarm B is delayed or could not be reported due to some reason(network problem, physical error and etc)."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_560a4f22","line":21,"in_reply_to":"1a430d35_7494e336","updated":"2017-02-07 01:47:00.000000000","message":"Done","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d2268c10c72ee08322519690d62fa81ee9729688","unresolved":false,"context_lines":[{"line_number":48,"context_line":"    e)  B-Vitrage causes C-Vitrage"},{"line_number":49,"context_line":""},{"line_number":50,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a \u0027equivalent\u0027 edge."},{"line_number":51,"context_line":"Currently, monitor alarms are identified as \u0027ALARM:nagios:node-1:alarm-A\u0027, so we have to make sure that alarm name must be unified in vitrage, not only in there own monitor."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"3. It may have more impact on performance because there are more vertex, and edges."},{"line_number":54,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_94994f71","line":51,"range":{"start_line":51,"start_character":44,"end_line":51,"end_character":45},"updated":"2017-02-05 02:51:20.000000000","message":"Use double backtick ``code``  for code quoting","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ee482b27a4b4caab1b1dceb9d4978dbddb2408cb","unresolved":false,"context_lines":[{"line_number":48,"context_line":"    e)  B-Vitrage causes C-Vitrage"},{"line_number":49,"context_line":""},{"line_number":50,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a \u0027equivalent\u0027 edge."},{"line_number":51,"context_line":"Currently, monitor alarms are identified as \u0027ALARM:nagios:node-1:alarm-A\u0027, so we have to make sure that alarm name must be unified in vitrage, not only in there own monitor."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"3. It may have more impact on performance because there are more vertex, and edges."},{"line_number":54,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_fb732413","line":51,"range":{"start_line":51,"start_character":44,"end_line":51,"end_character":45},"in_reply_to":"1a430d35_94994f71","updated":"2017-02-07 01:47:00.000000000","message":"Done","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d2268c10c72ee08322519690d62fa81ee9729688","unresolved":false,"context_lines":[{"line_number":57,"context_line":"Data model impact"},{"line_number":58,"context_line":"-----------------"},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"There would some impact on performance, because evaluator process may execute twice."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_b49e4b54","line":60,"range":{"start_line":60,"start_character":27,"end_line":60,"end_character":38},"updated":"2017-02-05 02:51:20.000000000","message":"Why in data model section?","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ee482b27a4b4caab1b1dceb9d4978dbddb2408cb","unresolved":false,"context_lines":[{"line_number":57,"context_line":"Data model impact"},{"line_number":58,"context_line":"-----------------"},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"There would some impact on performance, because evaluator process may execute twice."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_9b4060a7","line":60,"range":{"start_line":60,"start_character":27,"end_line":60,"end_character":38},"in_reply_to":"1a430d35_b49e4b54","updated":"2017-02-07 01:47:00.000000000","message":"Done","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d2268c10c72ee08322519690d62fa81ee9729688","unresolved":false,"context_lines":[{"line_number":126,"context_line":"Testing"},{"line_number":127,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":128,"context_line":""},{"line_number":129,"context_line":"Functinal tests and performance tests will be added."},{"line_number":130,"context_line":""},{"line_number":131,"context_line":""},{"line_number":132,"context_line":"Documentation Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_d4a3d79c","line":129,"range":{"start_line":129,"start_character":0,"end_line":129,"end_character":9},"updated":"2017-02-05 02:51:20.000000000","message":"\u003d\u003e Functional","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"ee482b27a4b4caab1b1dceb9d4978dbddb2408cb","unresolved":false,"context_lines":[{"line_number":126,"context_line":"Testing"},{"line_number":127,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":128,"context_line":""},{"line_number":129,"context_line":"Functinal tests and performance tests will be added."},{"line_number":130,"context_line":""},{"line_number":131,"context_line":""},{"line_number":132,"context_line":"Documentation Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1a430d35_d84aeec4","line":129,"range":{"start_line":129,"start_character":0,"end_line":129,"end_character":9},"in_reply_to":"1a430d35_d4a3d79c","updated":"2017-02-07 01:47:00.000000000","message":"Done","commit_id":"a7530a0e446bd8f82dbccfd2d2093752b66f3aa4"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"8993f3a2d3cd2a442f565282b3355d2eabc4a2e5","unresolved":false,"context_lines":[{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction"},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"This feature will allow us to get the deduced alarm that are not reported as"},{"line_number":14,"context_line":"expected (lost or delayed)."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"1a430d35_c0a8fe22","line":13,"range":{"start_line":13,"start_character":0,"end_line":13,"end_character":51},"updated":"2017-02-08 08:02:05.000000000","message":"What about\n\n\"Add support for deducing alarms that ... in vitrage\"\n\nIt is not just \"get\" but also \"generating\" and \"handling\". So \"support\" could be a more precise description.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"07d075bc5e0a24e195927dd7cee21f8271dcba3f","unresolved":false,"context_lines":[{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/vitrage/+spec/alarm-deduction"},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"This feature will allow us to get the deduced alarm that are not reported as"},{"line_number":14,"context_line":"expected (lost or delayed)."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"fa31d9ce_c69ed362","line":13,"range":{"start_line":13,"start_character":0,"end_line":13,"end_character":51},"in_reply_to":"1a430d35_c0a8fe22","updated":"2017-02-21 06:31:06.000000000","message":"Done","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"8993f3a2d3cd2a442f565282b3355d2eabc4a2e5","unresolved":false,"context_lines":[{"line_number":20,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the"},{"line_number":21,"context_line":"deduced alarm must be a vitrage type alarm."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will"},{"line_number":24,"context_line":"raise an alarm that should be raised by an external monitor, before the"},{"line_number":25,"context_line":"external monitor raises it."},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":28,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"}],"source_content_type":"text/x-rst","patch_set":5,"id":"1a430d35_8bdff767","line":25,"range":{"start_line":23,"start_character":0,"end_line":25,"end_character":27},"updated":"2017-02-08 08:02:05.000000000","message":"This seems to be part of proposed change, not the problem itself.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"07d075bc5e0a24e195927dd7cee21f8271dcba3f","unresolved":false,"context_lines":[{"line_number":20,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the"},{"line_number":21,"context_line":"deduced alarm must be a vitrage type alarm."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will"},{"line_number":24,"context_line":"raise an alarm that should be raised by an external monitor, before the"},{"line_number":25,"context_line":"external monitor raises it."},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":28,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"}],"source_content_type":"text/x-rst","patch_set":5,"id":"da36d5c6_4b749a9b","line":25,"range":{"start_line":23,"start_character":0,"end_line":25,"end_character":27},"in_reply_to":"1a430d35_8bdff767","updated":"2017-02-21 06:31:06.000000000","message":"I made some changes to the description.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"8993f3a2d3cd2a442f565282b3355d2eabc4a2e5","unresolved":false,"context_lines":[{"line_number":28,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":29,"context_line":"B is delayed or could not be reported due to some reason(network problem,"},{"line_number":30,"context_line":"physical error and etc)."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Proposed change"},{"line_number":34,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":5,"id":"1a430d35_abc3d387","line":31,"updated":"2017-02-08 08:02:05.000000000","message":"Then what is the problem? Why we need to deduce alarm B immediately? What is the benefit of doing so?","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"07d075bc5e0a24e195927dd7cee21f8271dcba3f","unresolved":false,"context_lines":[{"line_number":28,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":29,"context_line":"B is delayed or could not be reported due to some reason(network problem,"},{"line_number":30,"context_line":"physical error and etc)."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Proposed change"},{"line_number":34,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":5,"id":"fa31d9ce_66860715","line":31,"in_reply_to":"1a430d35_abc3d387","updated":"2017-02-21 06:31:06.000000000","message":"1. We could get the root alarm at any time and could deal with \nthe problems quickly.\n2. If the alarm could not reported(such as network problem), \nwe could also deduced that the alarm should exit in the graph.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":18783,"name":"dongwenjuan","email":"dong.wenjuan@zte.com.cn","username":"dongwenjuan"},"change_message_id":"f79d2b6376efe972df644e8970e45a02f3dac15f","unresolved":false,"context_lines":[{"line_number":43,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":44,"context_line":"   a new transformer may be needed."},{"line_number":45,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":46,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":47,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":48,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"1a430d35_cd6f45fa","line":46,"updated":"2017-02-08 13:01:39.000000000","message":"How to check the deduced alarm exists in the graph when the reported alarm arrived?","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"07d075bc5e0a24e195927dd7cee21f8271dcba3f","unresolved":false,"context_lines":[{"line_number":43,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":44,"context_line":"   a new transformer may be needed."},{"line_number":45,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":46,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":47,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":48,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"fa31d9ce_c6ac138f","line":46,"in_reply_to":"1a430d35_cd6f45fa","updated":"2017-02-21 06:31:06.000000000","message":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``. \nIf we could get \u0027datasource type\u0027 from the templates and add \nit to the deduced alarm, then we could get the same alarm ID \nas the monitor alarm in the graph.\nSo, we could identify the two alarms as the same by the alarm ID.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"6af682bc9e3ac5c5338c5de84acc84319c26f364","unresolved":false,"context_lines":[{"line_number":43,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":44,"context_line":"   a new transformer may be needed."},{"line_number":45,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":46,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":47,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":48,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"da36d5c6_7ddc2446","line":46,"in_reply_to":"da36d5c6_d12bdce9","updated":"2017-02-25 03:57:43.000000000","message":"\"datasource\" + \"alarm name\" + \"resource type\" + \"resource id\", all these elements could identify a unique alarm, we could get these elements from the properties of every vertex. So we could view the two alarms as the same if they have the same value on these elements.","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":18783,"name":"dongwenjuan","email":"dong.wenjuan@zte.com.cn","username":"dongwenjuan"},"change_message_id":"f9ccf6bf1bf79331e1c5f3536a6bd96ed5c3610e","unresolved":false,"context_lines":[{"line_number":43,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":44,"context_line":"   a new transformer may be needed."},{"line_number":45,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":46,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":47,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":48,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"da36d5c6_d12bdce9","line":46,"in_reply_to":"fa31d9ce_c6ac138f","updated":"2017-02-21 08:32:41.000000000","message":"As far as i know, the vitrage_id will be changed to UUID, not a set of characters, see the BP: https://blueprints.launchpad.net/vitrage/+spec/standard-vitrage-id. Then how do you get that?","commit_id":"1e9cfe185d66d9d42d67d741ebdc61d7e6c11479"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"242e147848a2c0ad658c93e1b77f940f401e3269","unresolved":false,"context_lines":[{"line_number":35,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We could identify the raised alarm B by vitrage and the reported alarm B by"},{"line_number":38,"context_line":"other datasource as the same alarm. In other words, we could merge the two"},{"line_number":39,"context_line":"alarms. We could just replace the vitrage raised alarm to datasource reported"},{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_dccc7b6c","line":38,"range":{"start_line":38,"start_character":20,"end_line":38,"end_character":34},"updated":"2017-02-23 07:23:45.000000000","message":"This is ambiguous. Since the source of the alarms are different, we could not say they are the same. However, you may describe it as:\n\nWe may model the two alarms as an identical vertex in entity graph.","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"1d6b5c3774abc1d0f885d1f0986cec4228b3a562","unresolved":false,"context_lines":[{"line_number":35,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We could identify the raised alarm B by vitrage and the reported alarm B by"},{"line_number":38,"context_line":"other datasource as the same alarm. In other words, we could merge the two"},{"line_number":39,"context_line":"alarms. We could just replace the vitrage raised alarm to datasource reported"},{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_fdf014cb","line":38,"range":{"start_line":38,"start_character":20,"end_line":38,"end_character":34},"in_reply_to":"da36d5c6_dccc7b6c","updated":"2017-02-25 03:58:01.000000000","message":"Done","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"242e147848a2c0ad658c93e1b77f940f401e3269","unresolved":false,"context_lines":[{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_bc38a73e","line":43,"updated":"2017-02-23 07:23:45.000000000","message":"Not quite understanding it.\n\nDo you mean raise an alarm with other datasource than `vitrage` from scenario evaluator?","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"7a1f7b4907d77a6f90c29b4baf7eff835cd5972b","unresolved":false,"context_lines":[{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_bcd427bd","line":43,"in_reply_to":"da36d5c6_bc38a73e","updated":"2017-02-23 07:26:01.000000000","message":"Correction: raise an alarm from scenario evaluator but specify another datasource?","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"1d6b5c3774abc1d0f885d1f0986cec4228b3a562","unresolved":false,"context_lines":[{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_7d342449","line":43,"in_reply_to":"da36d5c6_bcd427bd","updated":"2017-02-25 03:58:01.000000000","message":"\"datasource\" + \"alarm name\" + \"resource type\" + \"resource id\", all these elements could identify a unique alarm.If we want model the two alarms(the raised alarm by vitrage and the reported alarm) as an identical vertex in entity graph, we have to save the \"datasource\" information. So we could raise an alarm from scenario evaluator but specify another datasource.","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"242e147848a2c0ad658c93e1b77f940f401e3269","unresolved":false,"context_lines":[{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_1c48d3cb","line":47,"range":{"start_line":47,"start_character":15,"end_line":47,"end_character":22},"updated":"2017-02-23 07:23:45.000000000","message":"typo of replace ?","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"1d6b5c3774abc1d0f885d1f0986cec4228b3a562","unresolved":false,"context_lines":[{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_5dc04050","line":47,"range":{"start_line":47,"start_character":15,"end_line":47,"end_character":22},"in_reply_to":"da36d5c6_1c48d3cb","updated":"2017-02-25 03:58:01.000000000","message":"Done","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"242e147848a2c0ad658c93e1b77f940f401e3269","unresolved":false,"context_lines":[{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_bc86e7d8","line":48,"updated":"2017-02-23 07:23:45.000000000","message":"The recover of alarm B is also a deduced not reported. Would it be appropriate to simply removed it?\n\nWe should consider the raise and recover of alarm in same manner.","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d37ab5f2ce4cb3a581c27cc99f658d51428fbd31","unresolved":false,"context_lines":[{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_60a1d845","line":48,"in_reply_to":"da36d5c6_7d98a46a","updated":"2017-02-25 08:19:19.000000000","message":"I\u0027m making a comparison to bullet 2. When alarm A is monitored, alarm B is raise by vitrage but with \u0027is_deduced\u0027 property. This give us a way to distinguish a reported alarm B with deduced alarm B.\n\nWhen alarm A is recovered, how do we mark the recovery of alarm B is also deduced, not reported?","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":24666,"name":"wang weiya","email":"wang.weiya@zte.com.cn","username":"wang.weiya"},"change_message_id":"1d6b5c3774abc1d0f885d1f0986cec4228b3a562","unresolved":false,"context_lines":[{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_7d98a46a","line":48,"in_reply_to":"da36d5c6_bc86e7d8","updated":"2017-02-25 03:58:01.000000000","message":"If alarm A causes alarm B, and no other alarms could cause alarm B, then we could deduced that if alarm A is recovered then alarm B would be recovered, too. It depends on the RCA models.","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"d37ab5f2ce4cb3a581c27cc99f658d51428fbd31","unresolved":false,"context_lines":[{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just repalce the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"},{"line_number":52,"context_line":"twice."}],"source_content_type":"text/x-rst","patch_set":6,"id":"da36d5c6_60b8f875","line":49,"updated":"2017-02-25 08:19:19.000000000","message":"In this case, we should convert the recovery of alarm B from deduced to reported.\n\nIn my point of view, deduced alarm and reported alarm have different confidence level and may trigger different action. For deduced alarm, we normally runs a diagnostic in system to confirm the fault. While for reported alarm, normally an immediate evacuation is required.","commit_id":"9340a1b0e1e8dfd8b7b28fa662b463b821d24926"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":17,"context_line":"Problem description"},{"line_number":18,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the"},{"line_number":21,"context_line":"deduced alarm must be a vitrage type alarm."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will"},{"line_number":24,"context_line":"raise an alarm that should be raised by an external monitor, before the"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_5eac918c","line":21,"range":{"start_line":20,"start_character":53,"end_line":21,"end_character":43},"updated":"2017-02-27 16:51:44.000000000","message":"I don\u0027t see it as a restriction, but as a way of specifying that this alarm was raised by Vitrage.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":17,"context_line":"Problem description"},{"line_number":18,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the"},{"line_number":21,"context_line":"deduced alarm must be a vitrage type alarm."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will"},{"line_number":24,"context_line":"raise an alarm that should be raised by an external monitor, before the"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_c808ab30","line":21,"range":{"start_line":20,"start_character":53,"end_line":21,"end_character":43},"in_reply_to":"da36d5c6_5eac918c","updated":"2017-02-28 08:00:00.000000000","message":"I Agree. It wasn\u0027t a restriction from our side, but the way we wanted to represent deduced entities.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"ad02e23931d60001e11dc8c2ccda7792318894ee","unresolved":false,"context_lines":[{"line_number":17,"context_line":"Problem description"},{"line_number":18,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"Vitrage already supports alarm deduction, but it has a restriction that the"},{"line_number":21,"context_line":"deduced alarm must be a vitrage type alarm."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Vitrage should support a special case of alarm deduction -  that Vitrage will"},{"line_number":24,"context_line":"raise an alarm that should be raised by an external monitor, before the"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_6e56e707","line":21,"range":{"start_line":20,"start_character":53,"end_line":21,"end_character":43},"in_reply_to":"da36d5c6_c808ab30","updated":"2017-02-28 08:23:05.000000000","message":"I agree.\n\nThe `type` property in vitrage represented source of the alarm. This is a design, not restriction :-)","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":26,"context_line":"reason(network problem, physical error and etc), the RCA would be incomplete"},{"line_number":27,"context_line":"and the Root Cause may be incorrect."},{"line_number":28,"context_line":""},{"line_number":29,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":30,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":31,"context_line":"B is delayed or could not be reported due to some reason."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_9e8d292d","line":31,"range":{"start_line":29,"start_character":0,"end_line":31,"end_character":57},"updated":"2017-02-27 16:51:44.000000000","message":"Isn\u0027t it the opposite direction?\nWhen alarm B is reported, you would like to raise alarm A as if it was raised by the external monitor. Right?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"5fc4441fc8e2fa841ae890bf6af577fdf793cad6","unresolved":false,"context_lines":[{"line_number":26,"context_line":"reason(network problem, physical error and etc), the RCA would be incomplete"},{"line_number":27,"context_line":"and the Root Cause may be incorrect."},{"line_number":28,"context_line":""},{"line_number":29,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":30,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":31,"context_line":"B is delayed or could not be reported due to some reason."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_c457af24","line":31,"range":{"start_line":29,"start_character":0,"end_line":31,"end_character":57},"in_reply_to":"da36d5c6_0eb64b3d","updated":"2017-02-28 11:30:53.000000000","message":"+1 for more concrete examples","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"ad02e23931d60001e11dc8c2ccda7792318894ee","unresolved":false,"context_lines":[{"line_number":26,"context_line":"reason(network problem, physical error and etc), the RCA would be incomplete"},{"line_number":27,"context_line":"and the Root Cause may be incorrect."},{"line_number":28,"context_line":""},{"line_number":29,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":30,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":31,"context_line":"B is delayed or could not be reported due to some reason."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_0eb64b3d","line":31,"range":{"start_line":29,"start_character":0,"end_line":31,"end_character":57},"in_reply_to":"da36d5c6_4b2af57f","updated":"2017-02-28 08:23:05.000000000","message":"I think @Weiya is talking about causal relationship between **two reported** alarm. Correct me if wrong.\n\nMaybe we should use more concrete example here, e.g. `nagios-power-fault` causes `nagios-host-fault` so it won\u0027t cause unnecessary misunderstanding.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":26,"context_line":"reason(network problem, physical error and etc), the RCA would be incomplete"},{"line_number":27,"context_line":"and the Root Cause may be incorrect."},{"line_number":28,"context_line":""},{"line_number":29,"context_line":"For example, there is a known fault model: alarm A causes an alarm B. When"},{"line_number":30,"context_line":"alarm A is reported, we hope we can deduced the alarm B at once, even if alarm"},{"line_number":31,"context_line":"B is delayed or could not be reported due to some reason."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_4b2af57f","line":31,"range":{"start_line":29,"start_character":0,"end_line":31,"end_character":57},"in_reply_to":"da36d5c6_9e8d292d","updated":"2017-02-28 08:00:00.000000000","message":"From What I understood, I agree with Ifat. You talked about the other side of deducing alarms, not?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":34,"context_line":"Proposed change"},{"line_number":35,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"We may model the two alarms(the raised alarm B by vitrage and the reported alarm B)"},{"line_number":38,"context_line":"as an identical vertex in entity graph. In other words, we could merge the two"},{"line_number":39,"context_line":"alarms. We could just replace the vitrage raised alarm to datasource reported"},{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_6ba41908","line":40,"range":{"start_line":37,"start_character":0,"end_line":40,"end_character":6},"updated":"2017-02-28 08:00:00.000000000","message":"There is no need to merge the alarm. If the new deduced alarm B will have the same type (lets say Zabbix) and the same type (lets say cpu_overload as in zabbix) then the processor will recognize they are the same. So from my prospection, there is no problem with this kind of alarm correlation that Vitrage already handles.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":39,"context_line":"alarms. We could just replace the vitrage raised alarm to datasource reported"},{"line_number":40,"context_line":"alarm."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_7ec0752b","line":42,"range":{"start_line":42,"start_character":3,"end_line":42,"end_character":80},"updated":"2017-02-27 16:51:44.000000000","message":"I\u0027m not sure this is the wanted behaviour. As a user, I would like to know if the alarm was raised by Vitrage or by Zabbix.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_7eeeb5b5","line":44,"range":{"start_line":44,"start_character":52,"end_line":44,"end_character":62},"updated":"2017-02-27 16:51:44.000000000","message":"All Vitrage alarms are deduced. We need this flag only if we mark the alarm as a \"Zabbix\" alarm.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_cb47c51b","line":44,"range":{"start_line":44,"start_character":52,"end_line":44,"end_character":62},"in_reply_to":"da36d5c6_7eeeb5b5","updated":"2017-02-28 08:00:00.000000000","message":"In this case if the deduced alarm is \"Zabbix\" that w will mark it as \"is_deduced\" and when the real alarm will arrive from Zabbix itself then this property should be removed because it is no longer \"is_deduced\".\nI am not sure a new trasnformer is needed this case.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"5fc4441fc8e2fa841ae890bf6af577fdf793cad6","unresolved":false,"context_lines":[{"line_number":41,"context_line":""},{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_07fe2905","line":44,"range":{"start_line":44,"start_character":52,"end_line":44,"end_character":62},"in_reply_to":"da36d5c6_cb47c51b","updated":"2017-02-28 11:30:53.000000000","message":"And what if the real alarm never arrives? will we keep showing a \"fake\" zabbix alarm? showing it as a vitrage deduced alarm makes sense (that\u0027s what Vitrage is all about), but I don\u0027t like marking it as \"zabbix\"","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":18783,"name":"dongwenjuan","email":"dong.wenjuan@zte.com.cn","username":"dongwenjuan"},"change_message_id":"78340dd03fb92658e93ddde0017ea7d50448d253","unresolved":false,"context_lines":[{"line_number":42,"context_line":"1. Templates should specify the datasource type when the action is \u0027raise-alarm\u0027"},{"line_number":43,"context_line":"   (Default type is vitrage)."},{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_110ea2e4","line":45,"updated":"2017-02-27 02:55:09.000000000","message":"I think the `EvaluatorEventTransformer` already exits.  We don\u0027t need to add a new one.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_be982d1d","line":47,"range":{"start_line":47,"start_character":15,"end_line":47,"end_character":33},"updated":"2017-02-27 16:51:44.000000000","message":"Maybe instead of replacing, we should unite the content?\nIf we replace the vertex, all deducing will be re-calculated, i.e. the deduced alarms will be deleted and new ones will be created for the new datasource alarm. I don\u0027t think this is the wanted behaviour.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":44,"context_line":"2. Alarms from evaluator_queue should be added the \u0027is_deduced\u0027 property, and"},{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_4b30359f","line":47,"range":{"start_line":47,"start_character":15,"end_line":47,"end_character":33},"in_reply_to":"da36d5c6_be982d1d","updated":"2017-02-28 08:00:00.000000000","message":"I agree, maybe we should just unite the data content? (and as I said before, remove the \"is_deduced\" property).\nOtherwise we should maybe add a flag in the processor that says if we want to replace the vertex if exists or update it.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":45,"context_line":"   a new transformer may be needed."},{"line_number":46,"context_line":"3. Alarms from mp_queue should first check if the vertex exists in the graph."},{"line_number":47,"context_line":"   If yes just replace the vertex. Otherwise, continue the original process."},{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"},{"line_number":52,"context_line":"twice."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_deb40184","line":49,"range":{"start_line":48,"start_character":0,"end_line":49,"end_character":76},"updated":"2017-02-27 16:51:44.000000000","message":"What do you mean by recovered?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":48,"context_line":"4. If Alarm A is recovered, then alarm B is deduced to be recovered too."},{"line_number":49,"context_line":"   If alarm B recovered message is received later, just discard the message."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"There would some impact on performance, because evaluator process may execute"},{"line_number":52,"context_line":"twice."},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"Alternatives"},{"line_number":55,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_520f252e","line":52,"range":{"start_line":51,"start_character":0,"end_line":52,"end_character":6},"updated":"2017-02-27 16:51:44.000000000","message":"1. Please move it to a \"Performance Impact\" section at the end. \n2. I think that evaluating the same vertex twice is not a big deal, since most likely there will be only a few alarms. \n3. I\u0027m more concerned about the fact that for each alarm you will need to search the entity graph for a set of properties to see if an equivalent alarm already exists.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19233,"name":"Elisha Rosensweig","email":"elisha@chorus.ai","username":"elisha.rosensweig"},"change_message_id":"60962268bdd7e73478952e6b363e6982d9eac998","unresolved":false,"context_lines":[{"line_number":54,"context_line":"Alternatives"},{"line_number":55,"context_line":"------------"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"There is an alternative: have both vitrage alarm and the monitor alarm, and"},{"line_number":58,"context_line":"connect them with a \u0027equivalent\u0027 edge. We should discuss the advantages and"},{"line_number":59,"context_line":"disadvantages of each solution."},{"line_number":60,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_b8365ce5","line":57,"updated":"2017-02-28 12:56:49.000000000","message":"Note that this solution would support a much broader set of features. For example, supporting alarm aggregation, where we indicate that a set of alarms is actually one \"aggregate\" alarm. As such, it is a more generic solution than the one proposed above, which would only work for Vitrage (I\u0027m assuming if Nagios and Zabbix both raised an alarm, we would not want to have to decide which one is the \"real\" alarm)","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"31156377d5d00df537ebf293e6d80e6994281fb3","unresolved":false,"context_lines":[{"line_number":54,"context_line":"Alternatives"},{"line_number":55,"context_line":"------------"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"There is an alternative: have both vitrage alarm and the monitor alarm, and"},{"line_number":58,"context_line":"connect them with a \u0027equivalent\u0027 edge. We should discuss the advantages and"},{"line_number":59,"context_line":"disadvantages of each solution."},{"line_number":60,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ba2be162_d7dce648","line":57,"in_reply_to":"da36d5c6_b8365ce5","updated":"2017-02-28 15:08:40.000000000","message":"I agree. Thinking about it again, I think the \"alternative\" solution is better.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":61,"context_line":"In this alternative, all information are contained in the graph without any"},{"line_number":62,"context_line":"missing. But it may be more complex:"},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"1. The templates would be more complex, because vitrage alarm and the monitor"},{"line_number":65,"context_line":"   alarm are two different alarms in the graph."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"For example, alarm A causes alarm B, and alarm B causes alarm C, we have to"},{"line_number":68,"context_line":"specify these relations:"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_12ce8db4","line":65,"range":{"start_line":64,"start_character":3,"end_line":65,"end_character":47},"updated":"2017-02-27 16:51:44.000000000","message":"I agree","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19122,"name":"Alexey Weyl","email":"aloshaweyl@gmail.com","username":"alexey_weyl"},"change_message_id":"d16a7c18114d32b6087328384367fe126cce8baa","unresolved":false,"context_lines":[{"line_number":61,"context_line":"In this alternative, all information are contained in the graph without any"},{"line_number":62,"context_line":"missing. But it may be more complex:"},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"1. The templates would be more complex, because vitrage alarm and the monitor"},{"line_number":65,"context_line":"   alarm are two different alarms in the graph."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"For example, alarm A causes alarm B, and alarm B causes alarm C, we have to"},{"line_number":68,"context_line":"specify these relations:"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_6b8179fb","line":65,"range":{"start_line":64,"start_character":3,"end_line":65,"end_character":47},"in_reply_to":"da36d5c6_12ce8db4","updated":"2017-02-28 08:00:00.000000000","message":"I agree as well.\nI think the first option sounds the more \"Vitrage\" way to do it.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19233,"name":"Elisha Rosensweig","email":"elisha@chorus.ai","username":"elisha.rosensweig"},"change_message_id":"60962268bdd7e73478952e6b363e6982d9eac998","unresolved":false,"context_lines":[{"line_number":61,"context_line":"In this alternative, all information are contained in the graph without any"},{"line_number":62,"context_line":"missing. But it may be more complex:"},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"1. The templates would be more complex, because vitrage alarm and the monitor"},{"line_number":65,"context_line":"   alarm are two different alarms in the graph."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"For example, alarm A causes alarm B, and alarm B causes alarm C, we have to"},{"line_number":68,"context_line":"specify these relations:"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_ddda2217","line":65,"range":{"start_line":64,"start_character":3,"end_line":65,"end_character":47},"in_reply_to":"da36d5c6_6b8179fb","updated":"2017-02-28 12:56:49.000000000","message":"I disagree. see more comments below.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":72,"context_line":"    d)  B causes C-Vitrage"},{"line_number":73,"context_line":"    e)  B-Vitrage causes C-Vitrage"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a"},{"line_number":76,"context_line":"   \u0027equivalent\u0027 edge."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``,"},{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_92e35d33","line":76,"range":{"start_line":75,"start_character":3,"end_line":76,"end_character":21},"updated":"2017-02-27 16:51:44.000000000","message":"Why is it harder than merging the vertices?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"31156377d5d00df537ebf293e6d80e6994281fb3","unresolved":false,"context_lines":[{"line_number":72,"context_line":"    d)  B causes C-Vitrage"},{"line_number":73,"context_line":"    e)  B-Vitrage causes C-Vitrage"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a"},{"line_number":76,"context_line":"   \u0027equivalent\u0027 edge."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``,"},{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ba2be162_579f165f","line":76,"range":{"start_line":75,"start_character":3,"end_line":76,"end_character":21},"in_reply_to":"da36d5c6_3d2fbef7","updated":"2017-02-28 15:08:40.000000000","message":"But what if the existing alarms are A, B and C, but not B-Vitrage? you must add a rule for A causes B","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19233,"name":"Elisha Rosensweig","email":"elisha@chorus.ai","username":"elisha.rosensweig"},"change_message_id":"60962268bdd7e73478952e6b363e6982d9eac998","unresolved":false,"context_lines":[{"line_number":72,"context_line":"    d)  B causes C-Vitrage"},{"line_number":73,"context_line":"    e)  B-Vitrage causes C-Vitrage"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a"},{"line_number":76,"context_line":"   \u0027equivalent\u0027 edge."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``,"},{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_3d2fbef7","line":76,"range":{"start_line":75,"start_character":3,"end_line":76,"end_character":21},"in_reply_to":"da36d5c6_92e35d33","updated":"2017-02-28 12:56:49.000000000","message":"Why? It seems like the entity graph would have more connections, but I don\u0027t see why this is more \"difficult\".\n\nLooking at this example, we could specify as in the example above. Alternatively, if we assume that alarm B-Vitrage would precede B and alarm C-Vitrage would precede C (which the user would know from his own system), we could just do\n\n a) A causes B-Vitrage\n b) B-Vitrage causes C-Vitrage\n c) B equivalent B-Vitrage\n d) C equivalent C-Vitrage\n\nAnd have the RCA api traverse \"causes\" and \"equivalent\" edges","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":75,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a"},{"line_number":76,"context_line":"   \u0027equivalent\u0027 edge."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``,"},{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"},{"line_number":80,"context_line":"in there own monitor."},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"3. It may have more impact on performance because there are more vertex, and"},{"line_number":83,"context_line":"   edges."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_728fc975","line":80,"range":{"start_line":78,"start_character":0,"end_line":80,"end_character":21},"updated":"2017-02-27 16:51:44.000000000","message":"This is about to be changed in a short time, when Vitrage ID will be changed to UUID","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19233,"name":"Elisha Rosensweig","email":"elisha@chorus.ai","username":"elisha.rosensweig"},"change_message_id":"60962268bdd7e73478952e6b363e6982d9eac998","unresolved":false,"context_lines":[{"line_number":75,"context_line":"2. It would be hard to connect vitrage alarm and the monitor alarm with a"},{"line_number":76,"context_line":"   \u0027equivalent\u0027 edge."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Currently, monitor alarms are identified as ``ALARM:nagios:node-1:alarm-A``,"},{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"},{"line_number":80,"context_line":"in there own monitor."},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"3. It may have more impact on performance because there are more vertex, and"},{"line_number":83,"context_line":"   edges."}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_3d7dfee7","line":80,"range":{"start_line":78,"start_character":0,"end_line":80,"end_character":21},"in_reply_to":"da36d5c6_728fc975","updated":"2017-02-28 12:56:49.000000000","message":"Even after the change of adding a uuid we would still never have two identical alarms from the same source (e.g., vitrage, nagios) on the same resource in parallel. \n\nAs for name uniqueness - in general, you would expect that the same alarm name would reflect the same alarm - no?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"},{"line_number":80,"context_line":"in there own monitor."},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"3. It may have more impact on performance because there are more vertex, and"},{"line_number":83,"context_line":"   edges."},{"line_number":84,"context_line":""},{"line_number":85,"context_line":""},{"line_number":86,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_b20301be","line":83,"range":{"start_line":82,"start_character":0,"end_line":83,"end_character":9},"updated":"2017-02-27 16:51:44.000000000","message":"I agree. And you can also add #4, that it should require UI enhancements, because if we display two edges with an \"equivalent\" edge, the graph will look too complex","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19233,"name":"Elisha Rosensweig","email":"elisha@chorus.ai","username":"elisha.rosensweig"},"change_message_id":"60962268bdd7e73478952e6b363e6982d9eac998","unresolved":false,"context_lines":[{"line_number":79,"context_line":"so we have to make sure that alarm name must be unified in vitrage, not only"},{"line_number":80,"context_line":"in there own monitor."},{"line_number":81,"context_line":""},{"line_number":82,"context_line":"3. It may have more impact on performance because there are more vertex, and"},{"line_number":83,"context_line":"   edges."},{"line_number":84,"context_line":""},{"line_number":85,"context_line":""},{"line_number":86,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_9dda6ae8","line":83,"range":{"start_line":82,"start_character":0,"end_line":83,"end_character":9},"in_reply_to":"da36d5c6_b20301be","updated":"2017-02-28 12:56:49.000000000","message":"I don\u0027t see where the performance impacts are so evident here. Having an additional vertex per-usecase does not seem so much heavier than the current situation. Once we move to a persistent graph DB, and considering that we plan on dealing with history etc, we will anyway need to have support for larger and more complex graphs.\n\nAs for UI - yes, adding more complex relationships in the entity graph will require more sophisticated APIs and UI. But we know this is already the case from just dealing with large-scale entity graphs (again, think aggregation), so I doubt we will be able to avoid this regardless.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":101,"context_line":""},{"line_number":102,"context_line":"Templates should add \u0027datasource type\u0027 field when raise alarm, it is optional"},{"line_number":103,"context_line":"and default value is \u0027vitrage\u0027."},{"line_number":104,"context_line":"Old templates can work well in new version, but there will be some problem when"},{"line_number":105,"context_line":"new templates work in old version."},{"line_number":106,"context_line":""},{"line_number":107,"context_line":""},{"line_number":108,"context_line":"Other end user impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_521c8513","line":105,"range":{"start_line":104,"start_character":0,"end_line":105,"end_character":34},"updated":"2017-02-27 16:51:44.000000000","message":"Why should there be a problem if the parameter is optional?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":113,"context_line":""},{"line_number":114,"context_line":""},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"Deployer impact"},{"line_number":117,"context_line":"---------------"},{"line_number":118,"context_line":""},{"line_number":119,"context_line":"None"},{"line_number":120,"context_line":""},{"line_number":121,"context_line":""},{"line_number":122,"context_line":"Developer impact"},{"line_number":123,"context_line":"----------------"},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"None"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"Horizon impact"},{"line_number":129,"context_line":"--------------"},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"None"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_d2a4f5ba","line":131,"range":{"start_line":116,"start_character":0,"end_line":131,"end_character":4},"updated":"2017-02-27 16:51:44.000000000","message":"You can remove impacts that are None (assuming everyone agrees that they are indeed None)","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"d3e73a9b4dce89a985af12b104d7bfb1633c191f","unresolved":false,"context_lines":[{"line_number":113,"context_line":""},{"line_number":114,"context_line":""},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"Deployer impact"},{"line_number":117,"context_line":"---------------"},{"line_number":118,"context_line":""},{"line_number":119,"context_line":"None"},{"line_number":120,"context_line":""},{"line_number":121,"context_line":""},{"line_number":122,"context_line":"Developer impact"},{"line_number":123,"context_line":"----------------"},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"None"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"Horizon impact"},{"line_number":129,"context_line":"--------------"},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"None"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ba2be162_2769268c","line":131,"range":{"start_line":116,"start_character":0,"end_line":131,"end_character":4},"in_reply_to":"ba2be162_58e37005","updated":"2017-03-02 13:21:14.000000000","message":"OK","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":21414,"name":"Yujun Zhang","email":"zhang.yujunz@zte.com.cn","username":"yujunz"},"change_message_id":"ad08bee34ed448e65526ec62a5a718484e63bea7","unresolved":false,"context_lines":[{"line_number":113,"context_line":""},{"line_number":114,"context_line":""},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"Deployer impact"},{"line_number":117,"context_line":"---------------"},{"line_number":118,"context_line":""},{"line_number":119,"context_line":"None"},{"line_number":120,"context_line":""},{"line_number":121,"context_line":""},{"line_number":122,"context_line":"Developer impact"},{"line_number":123,"context_line":"----------------"},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"None"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"Horizon impact"},{"line_number":129,"context_line":"--------------"},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"None"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":7,"id":"ba2be162_58e37005","line":131,"range":{"start_line":116,"start_character":0,"end_line":131,"end_character":4},"in_reply_to":"da36d5c6_d2a4f5ba","updated":"2017-03-02 07:00:49.000000000","message":"I would prefer to keep them to show that we have considered impacts on these parts and the conclusion is None.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":153,"context_line":"Dependencies"},{"line_number":154,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"None"},{"line_number":157,"context_line":""},{"line_number":158,"context_line":""},{"line_number":159,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_12020d70","line":156,"range":{"start_line":156,"start_character":0,"end_line":156,"end_character":4},"updated":"2017-02-27 16:51:44.000000000","message":"It might depend on the Vitrage ID refactoring. Not sure.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"e7b7b165ee316e398aa24d55c40abc8d800b4be2","unresolved":false,"context_lines":[{"line_number":165,"context_line":"Documentation Impact"},{"line_number":166,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Template document should be updated."},{"line_number":169,"context_line":""},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_b503ebbd","line":168,"range":{"start_line":168,"start_character":0,"end_line":168,"end_character":36},"updated":"2017-02-27 16:51:44.000000000","message":"After we agree on the design basics, I would like to see a detailed use case document that \"proves\" that it works. Something like: \n\nUse case 1:\n- we receive a \"VM connectivity\" (or whatever alarm example that you have) alarm from Zabbix\n- here is the template that is going to be executed\n- we raise \"Host connectivity\" deduced/suspect alarm \n- we raise/don\u0027t raise \"VM connectivity\" deduced alarm from the \"Host connectivity\" deduced alarm\n- we get \"Host connectivity\" alarm from Zabbix\n- here is the relevant template (or maybe it is part of the previous template)\n- we replace/merge/link the \"Host connectivity\" vitrage alarm\n- we create/don\u0027t create other deduced alarms\n- this is how the final graph looks like, and now it is a stable state because ...\n\nUse case 2: same as #1 but the Zabbix alarm is never retrieved. Will the graph be stable? will everything else behave the same? or maybe you don\u0027t want it to behave the same?\n\nUse case 3: I believe you mentioned in the past a case where there could be two possible causes for alarm X, and you are not sure which of them occurred. What should be the behaviour then?\n\netc.\n\nI\u0027m not sure if these use cases should be part of the blueprint, or reside in a separate document under vitrage/doc. But I think they are very important for understanding the full flow.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"c4d188d0976be2e36e6ed69827629ecc1cb4bf14","unresolved":false,"context_lines":[{"line_number":165,"context_line":"Documentation Impact"},{"line_number":166,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Template document should be updated."},{"line_number":169,"context_line":""},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_3d8c1e35","line":168,"range":{"start_line":168,"start_character":0,"end_line":168,"end_character":36},"in_reply_to":"da36d5c6_87c3f989","updated":"2017-02-28 12:34:18.000000000","message":"Two more use cases:\n5. The deduced alarm is raised. Then the real alarm is raised. Then the real alarm goes down. What should happen to the deduced alarm?\n6. The real alarm is raised before the deduced alarm. What should happen?","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"},{"author":{"_account_id":19159,"name":"Ifat Afek","email":"ifat.afek@nokia.com","username":"ifat_afek"},"change_message_id":"5fc4441fc8e2fa841ae890bf6af577fdf793cad6","unresolved":false,"context_lines":[{"line_number":165,"context_line":"Documentation Impact"},{"line_number":166,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":167,"context_line":""},{"line_number":168,"context_line":"Template document should be updated."},{"line_number":169,"context_line":""},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"References"}],"source_content_type":"text/x-rst","patch_set":7,"id":"da36d5c6_87c3f989","line":168,"range":{"start_line":168,"start_character":0,"end_line":168,"end_character":36},"in_reply_to":"da36d5c6_b503ebbd","updated":"2017-02-28 11:30:53.000000000","message":"There is another use case - #4:\nWe create a vitrage alarm on the host. Then a Zabbix alarm arrives. Then a Nagios alarm with the same name arrives. What do we do?\n\nThis use case is more complex, and is not likely to occur, so I don\u0027t think we must solve it now. But let\u0027s write it down and make sure the behaviour in this case is reasonable.","commit_id":"de976f9299f7dccef4e423fbb335b8130610c78a"}]}
