)]}'
{"specs/mitaka/timeseries-dispatcher.rst":[{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","unresolved":false,"context_lines":[{"line_number":12,"context_line":""},{"line_number":13,"context_line":"The goal of this blueprint is to add a new dispatcher for timeseries databases."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"The first working prototype is done for InfluxDB, other storage drivers can be "},{"line_number":16,"context_line":"added later (OpenTSDB, KairosDB)."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Problem description"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_6288f5c5","line":15,"updated":"2015-08-25 22:06:06.000000000","message":"remove extra space","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","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":"* Send efficiently samples to a timeseries database without passing through a "},{"line_number":22,"context_line":"  third-party system"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"* Use the indexes of the timeseries database"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_228e7dd0","line":21,"updated":"2015-08-25 22:06:06.000000000","message":"ditto","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"89c5353773347a5630d4d2afd558568ca35e844d","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"* Use the indexes of the timeseries database"},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"* Support multiple timeseries databases: InfluxDB, OpenTSDB, KairosDB, Gnocchi?"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Proposed change"},{"line_number":29,"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":1,"id":"fa1b9901_21220253","line":26,"updated":"2015-08-25 19:52:34.000000000","message":"is this a abstract layer or a dispatcher? if this is a dispatcher, why can\u0027t you just use http dispather and post data to influxdb?","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17784,"name":"Frédéric Guillot","email":"fguillot@internap.com","username":"fguillot"},"change_message_id":"ddfbe0e7c23c8d6d0475b0d465b76aed5213f328","unresolved":false,"context_lines":[{"line_number":23,"context_line":""},{"line_number":24,"context_line":"* Use the indexes of the timeseries database"},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"* Support multiple timeseries databases: InfluxDB, OpenTSDB, KairosDB, Gnocchi?"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Proposed change"},{"line_number":29,"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":1,"id":"fa1b9901_cdd81721","line":26,"in_reply_to":"fa1b9901_21220253","updated":"2015-08-25 21:31:28.000000000","message":"It\u0027s a dispatcher that use an abstract layer for the database driver, so we have a generic timeseries dispatcher that can abstract any timeseries system, influxdb or even gnocchi, then people can choose the solution that fit their needs.\n\nUsing the http dispatcher will require an intermediate system that receive samples and save them to influxdb. Better to listen on a message queue (with rabbitmq) if we choose this way...","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","unresolved":false,"context_lines":[{"line_number":28,"context_line":"Proposed change"},{"line_number":29,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"* Add new timeseries dispatcher that filter samples and save them to the "},{"line_number":32,"context_line":"  database through an abstracted timeserie storage"},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"* Add new config file \"timeseries.yaml\" that define the list of samples and tags"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_e2c2c5e8","line":31,"updated":"2015-08-25 22:06:06.000000000","message":"ditto","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","unresolved":false,"context_lines":[{"line_number":34,"context_line":"* Add new config file \"timeseries.yaml\" that define the list of samples and tags"},{"line_number":35,"context_line":"  to save in the database"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* New timeseries storage drivers for different databases (InfluxDB, OpenTSDB, "},{"line_number":38,"context_line":"  KairosDB)"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_82b3c999","line":37,"updated":"2015-08-25 22:06:06.000000000","message":"ditto","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"99e21e65c5abd20a05a61f465b6e9b749f12b85c","unresolved":false,"context_lines":[{"line_number":35,"context_line":"  to save in the database"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* New timeseries storage drivers for different databases (InfluxDB, OpenTSDB, "},{"line_number":38,"context_line":"  KairosDB)"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"Alternatives"},{"line_number":41,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_92ea3e0e","line":38,"updated":"2015-08-25 20:41:23.000000000","message":"Are we sure we can abstract to that level?\n\nThat means the dispatcher is a driver of the collector and now it has its own driver… why not but I\u0027d rather be sure it\u0027s possible.","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","unresolved":false,"context_lines":[{"line_number":40,"context_line":"Alternatives"},{"line_number":41,"context_line":"------------"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Gnocchi can already be used as timeseries database but it doesn\u0027t necessary fit "},{"line_number":44,"context_line":"the need and the scalability of everybody. Most of the time features included in"},{"line_number":45,"context_line":"existing timeseries databases are enough."},{"line_number":46,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_a2b84d76","line":43,"updated":"2015-08-25 22:06:06.000000000","message":"ditto","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"89c5353773347a5630d4d2afd558568ca35e844d","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Gnocchi can already be used as timeseries database but it doesn\u0027t necessary fit "},{"line_number":44,"context_line":"the need and the scalability of everybody. Most of the time features included in"},{"line_number":45,"context_line":"existing timeseries databases are enough."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_81b8ae76","line":45,"updated":"2015-08-25 19:52:34.000000000","message":"is the structure in Gnocchi too rigid? this current proposal seems to lack standardisation. from what is described, the proposal is to dump the data in to a tsdb and the user must now figure a way to consume it from the db. is there a specific model to ensure consistency when data changes?","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17784,"name":"Frédéric Guillot","email":"fguillot@internap.com","username":"fguillot"},"change_message_id":"ddfbe0e7c23c8d6d0475b0d465b76aed5213f328","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Gnocchi can already be used as timeseries database but it doesn\u0027t necessary fit "},{"line_number":44,"context_line":"the need and the scalability of everybody. Most of the time features included in"},{"line_number":45,"context_line":"existing timeseries databases are enough."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_3f3c5ac7","line":45,"in_reply_to":"fa1b9901_81b8ae76","updated":"2015-08-25 21:31:28.000000000","message":"At the moment, there is no model. The only information sent to the tsdb is the sample name, the counter value, timestamp and optionnally some tags.\n\nExact, the user have to read the data from the db. I sent another blueprint yesterday with new api endpoints to solve this issue but apparently there is no way to add new stuff in the api.","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"99e21e65c5abd20a05a61f465b6e9b749f12b85c","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Gnocchi can already be used as timeseries database but it doesn\u0027t necessary fit "},{"line_number":44,"context_line":"the need and the scalability of everybody. Most of the time features included in"},{"line_number":45,"context_line":"existing timeseries databases are enough."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_f2b85a13","line":45,"in_reply_to":"fa1b9901_81b8ae76","updated":"2015-08-25 20:41:23.000000000","message":"I think there is a use case where you don\u0027t want to open the data to the user and just do things with the data directly without ensuring any compat from one Ceilometer deployment to the other.","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"89c5353773347a5630d4d2afd558568ca35e844d","unresolved":false,"context_lines":[{"line_number":82,"context_line":""},{"line_number":83,"context_line":"::"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"  - storage.containers.objects # meter name to store in the database"},{"line_number":86,"context_line":"      - project_id # tag to attach to the datapoint"},{"line_number":87,"context_line":"      - resource_id"},{"line_number":88,"context_line":"  - network.outgoing.bytes"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_21a522a1","line":85,"updated":"2015-08-25 19:52:34.000000000","message":"how is this defined? who defines it?","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":17120,"name":"Manjeet Singh Bhatia","email":"manjeet.s.bhatia@intel.com","username":"manjeets"},"change_message_id":"2fb30fcb4d254d064c5814f11f762f97211e0b34","unresolved":false,"context_lines":[{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  [DEFAULT]"},{"line_number":98,"context_line":"  dispatcher \u003d timeseries"},{"line_number":99,"context_line":"   "},{"line_number":100,"context_line":"  [timeserie]"},{"line_number":101,"context_line":"  metering_connection \u003d influxdb://root:root@localhost/ceilometer"},{"line_number":102,"context_line":"  metering_policy \u003d /etc/ceilometer/timeseries.yaml"}],"source_content_type":"text/x-rst","patch_set":1,"id":"fa1b9901_4204d148","line":99,"updated":"2015-08-25 22:06:06.000000000","message":"remove extra space","commit_id":"c7f8ef822344ba6705fdcbe163ffa1482ce8c5fb"},{"author":{"_account_id":13997,"name":"SHIGEMATSU Mitsuhiro","email":"shigematsu.mitsuhiro@lab.ntt.co.jp","username":"pshige"},"change_message_id":"2bf04143c02fd5122eb330f3f2ad80cc47d41087","unresolved":false,"context_lines":[{"line_number":69,"context_line":""},{"line_number":70,"context_line":"* The exact same test was done with Gnocchi and InfluxDB resulted in the Gnocchi"},{"line_number":71,"context_line":"  processes crashing after sending 38000 samples and the disk usage of the"},{"line_number":72,"context_line":"  InfluxDb data folder was only 1.8MB with the same archive policy"},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"* Then we tested our basic Ceilometer dispatcher that was sending everything"},{"line_number":75,"context_line":"  directly to InfluxDB and we didn\u0027t observe any crashes. We were able to run"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da20952f_5069db8a","line":72,"updated":"2015-09-05 03:22:30.000000000","message":"s/InfluxDb/InfluxDB/","commit_id":"198b2e7cecdb47dcdfe7fa8c3da4cb7a152add42"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"167b544e88f1a0190567c4784303c54c1d702929","unresolved":false,"context_lines":[{"line_number":88,"context_line":"* Add a new time-series storage driver for InfluxDB"},{"line_number":89,"context_line":""},{"line_number":90,"context_line":"* Add a new configuration file \"timeseries.yaml\" that defines the list of"},{"line_number":91,"context_line":"  metrics to save in the database"},{"line_number":92,"context_line":""},{"line_number":93,"context_line":""},{"line_number":94,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_2d0c4840","line":91,"updated":"2015-11-10 21:40:28.000000000","message":"what\u0027s the api for this?","commit_id":"198b2e7cecdb47dcdfe7fa8c3da4cb7a152add42"},{"author":{"_account_id":18062,"name":"Maxime Belanger","email":"mbelanger@internap.com","username":"mbelanger"},"change_message_id":"08b6235ea3c74606c0fcc46b7138d10193b1fcb9","unresolved":false,"context_lines":[{"line_number":88,"context_line":"* Add a new time-series storage driver for InfluxDB"},{"line_number":89,"context_line":""},{"line_number":90,"context_line":"* Add a new configuration file \"timeseries.yaml\" that defines the list of"},{"line_number":91,"context_line":"  metrics to save in the database"},{"line_number":92,"context_line":""},{"line_number":93,"context_line":""},{"line_number":94,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":2,"id":"da85f559_6ea8e18d","line":91,"in_reply_to":"da85f559_2d0c4840","updated":"2016-03-10 21:56:55.000000000","message":"Do you mean that we can propose a new API for this in Mitaka? It was our original intention to do so.","commit_id":"198b2e7cecdb47dcdfe7fa8c3da4cb7a152add42"}]}