)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"89e065ee02cbff0ce8b9793b54e4ae30e0c65792","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This spec outlines approach for integrating"},{"line_number":10,"context_line":"Monasca as a ceilometer data collection backend."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"DocImpact"},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"blueprint monasca-backend"},{"line_number":15,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":3,"id":"9a8ffd7b_ffef14ae","line":12,"updated":"2015-11-25 09:29:40.000000000","message":"I don\u0027t think this is relevant here.","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"8c2a060a394f3dd24fbb31c92bc5ac3e60393e77","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This spec outlines approach for integrating"},{"line_number":10,"context_line":"Monasca as a ceilometer data collection backend."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"DocImpact"},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"blueprint monasca-backend"},{"line_number":15,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":3,"id":"9a8ffd7b_2da9df7b","line":12,"in_reply_to":"9a8ffd7b_ffef14ae","updated":"2015-11-25 15:28:11.000000000","message":"agreed","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"}],"specs/mitaka/monasca-backend.rst":[{"author":{"_account_id":12512,"name":"Deklan Dieterly","email":"deklan.dieterly@hpe.com","username":"deklan"},"change_message_id":"3ee98b54cae2ed3b503d7fb626916d271d6607e6","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Ceilometer monitors and collects data from OpenStack via notifications sent"},{"line_number":20,"context_line":"from existing services or by polling the infrastructure. Ceilometer has historically"},{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_b4930afa","line":21,"updated":"2015-11-24 20:43:49.000000000","message":"\"where\" -\u003e \"whereas\"","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Ceilometer monitors and collects data from OpenStack via notifications sent"},{"line_number":20,"context_line":"from existing services or by polling the infrastructure. Ceilometer has historically"},{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_ba5531c7","line":21,"in_reply_to":"9a8ffd7b_b4930afa","updated":"2015-11-24 21:16:10.000000000","message":"This is FUD and completely irrelevant into a spec.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":18,"context_line":""},{"line_number":19,"context_line":"Ceilometer monitors and collects data from OpenStack via notifications sent"},{"line_number":20,"context_line":"from existing services or by polling the infrastructure. Ceilometer has historically"},{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_d7680bbd","line":21,"in_reply_to":"9a8ffd7b_ba5531c7","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":20,"context_line":"from existing services or by polling the infrastructure. Ceilometer has historically"},{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_9a638db1","line":23,"updated":"2015-11-24 21:16:10.000000000","message":"Ceilometer","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":20,"context_line":"from existing services or by polling the infrastructure. Ceilometer has historically"},{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_f7650fe1","line":23,"in_reply_to":"9a8ffd7b_9a638db1","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":12512,"name":"Deklan Dieterly","email":"deklan.dieterly@hpe.com","username":"deklan"},"change_message_id":"3ee98b54cae2ed3b503d7fb626916d271d6607e6","unresolved":false,"context_lines":[{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."},{"line_number":27,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_d47ecead","line":24,"updated":"2015-11-24 20:43:49.000000000","message":"period instead of comma","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":21,"context_line":"had issues with scaling where Monasca is known to scale well."},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."},{"line_number":27,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_976283d7","line":24,"in_reply_to":"9a8ffd7b_d47ecead","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Batching of samples is an important feature to achieve high insert throughput. Most ceilometer"},{"line_number":29,"context_line":"backends do not support batching and the ones that do are for datapoints for the same meter,"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_ba6c519d","line":26,"updated":"2015-11-24 21:16:10.000000000","message":"FUD again. No data, nothing.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":23,"context_line":"When consuming large scale data, ceilometer encounters bottleneck while republishing"},{"line_number":24,"context_line":"of samples to rabbitMQ, In large deployments, rabbitMQ clustering has some limitations"},{"line_number":25,"context_line":"around the 20M-mark load [1]; this can slow down the queue performance and impact other"},{"line_number":26,"context_line":"services relying on the queue."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Batching of samples is an important feature to achieve high insert throughput. Most ceilometer"},{"line_number":29,"context_line":"backends do not support batching and the ones that do are for datapoints for the same meter,"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_57041b7d","line":26,"in_reply_to":"9a8ffd7b_ba6c519d","updated":"2015-11-24 22:59:48.000000000","message":"we have conducted extensive performance comparison testing between ceilometer w/mongo and ceilometer w/monasca and this is the outcome of those results. I will update the information in the spec.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":31,"context_line":"to support batching for cross-meter samples."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Monasca is a highly scalable monitoring solution and exposes a REST API for"},{"line_number":34,"context_line":"metric and measurement processing. which stores them into Kafka. Kafka is known to be performant"},{"line_number":35,"context_line":"at much higher loads that RMQ can support. This allows for a different publishing rate from the"},{"line_number":36,"context_line":"storage rate in Ceilometer collector, thereby increasing performance and optimization at the"},{"line_number":37,"context_line":"distinct layers."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_3addc142","line":34,"updated":"2015-11-24 21:16:10.000000000","message":"I find it weird to say \"store into Kafka\" since AFAIK Kafka is a transport system.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":12512,"name":"Deklan Dieterly","email":"deklan.dieterly@hpe.com","username":"deklan"},"change_message_id":"3ee98b54cae2ed3b503d7fb626916d271d6607e6","unresolved":false,"context_lines":[{"line_number":31,"context_line":"to support batching for cross-meter samples."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Monasca is a highly scalable monitoring solution and exposes a REST API for"},{"line_number":34,"context_line":"metric and measurement processing. which stores them into Kafka. Kafka is known to be performant"},{"line_number":35,"context_line":"at much higher loads that RMQ can support. This allows for a different publishing rate from the"},{"line_number":36,"context_line":"storage rate in Ceilometer collector, thereby increasing performance and optimization at the"},{"line_number":37,"context_line":"distinct layers."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_5f0e7735","line":34,"updated":"2015-11-24 20:43:49.000000000","message":"remove period and capitalize which","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":31,"context_line":"to support batching for cross-meter samples."},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"Monasca is a highly scalable monitoring solution and exposes a REST API for"},{"line_number":34,"context_line":"metric and measurement processing. which stores them into Kafka. Kafka is known to be performant"},{"line_number":35,"context_line":"at much higher loads that RMQ can support. This allows for a different publishing rate from the"},{"line_number":36,"context_line":"storage rate in Ceilometer collector, thereby increasing performance and optimization at the"},{"line_number":37,"context_line":"distinct layers."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_77a03ff5","line":34,"in_reply_to":"9a8ffd7b_3addc142","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":34,"context_line":"metric and measurement processing. which stores them into Kafka. Kafka is known to be performant"},{"line_number":35,"context_line":"at much higher loads that RMQ can support. This allows for a different publishing rate from the"},{"line_number":36,"context_line":"storage rate in Ceilometer collector, thereby increasing performance and optimization at the"},{"line_number":37,"context_line":"distinct layers."},{"line_number":38,"context_line":""},{"line_number":39,"context_line":"By integrating Monasca as a ceilometer backend, ceilometer can leverage monasca ability to batch"},{"line_number":40,"context_line":"datapoints and bypass re-publishing to RMQ. It also allows for interesting possibilities for new use"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_3ab4e1f0","line":37,"updated":"2015-11-24 21:16:10.000000000","message":"FUD unless you have something to back this up.\n\nThough if it\u0027s true, we should definitely build a Kafka transport for notifications then.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":39,"context_line":"By integrating Monasca as a ceilometer backend, ceilometer can leverage monasca ability to batch"},{"line_number":40,"context_line":"datapoints and bypass re-publishing to RMQ. It also allows for interesting possibilities for new use"},{"line_number":41,"context_line":"cases around metering and monitoring since Monasca is backend for monitoring metrics as well."},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Last but not the least, Monasca API also offers faster query response times and paging. Its found to"},{"line_number":44,"context_line":"be much faster than the ceilometer v2 API [2]. It offers an interesting option to consider integrating"},{"line_number":45,"context_line":"Monasca API as a ceilometer storage driver or as an alternative for the ceilometer v2 API."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_fa95f93f","line":42,"updated":"2015-11-24 21:16:10.000000000","message":"I don\u0027t think we need Monasca to use Kafka?","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":39,"context_line":"By integrating Monasca as a ceilometer backend, ceilometer can leverage monasca ability to batch"},{"line_number":40,"context_line":"datapoints and bypass re-publishing to RMQ. It also allows for interesting possibilities for new use"},{"line_number":41,"context_line":"cases around metering and monitoring since Monasca is backend for monitoring metrics as well."},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Last but not the least, Monasca API also offers faster query response times and paging. Its found to"},{"line_number":44,"context_line":"be much faster than the ceilometer v2 API [2]. It offers an interesting option to consider integrating"},{"line_number":45,"context_line":"Monasca API as a ceilometer storage driver or as an alternative for the ceilometer v2 API."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_77d59f41","line":42,"in_reply_to":"9a8ffd7b_fa95f93f","updated":"2015-11-24 22:59:48.000000000","message":"My apologies for  not clarifying this enough, we intend to use the Monasca-api in the publisher, i will update the spec to make it explicit.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Last but not the least, Monasca API also offers faster query response times and paging. Its found to"},{"line_number":44,"context_line":"be much faster than the ceilometer v2 API [2]. It offers an interesting option to consider integrating"},{"line_number":45,"context_line":"Monasca API as a ceilometer storage driver or as an alternative for the ceilometer v2 API."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_facad94e","line":45,"updated":"2015-11-24 21:16:10.000000000","message":"Ceilometer v2 API is already deprecated. We have Gnocchi, which should be even faster than Monasca, so it\u0027s FUD again.\n\nBut I\u0027m not specially pushing Gnocchi as it\u0027s also different than the Ceilometer v2 API, but it\u0027s just to point out how irrelevant this argument is here.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Last but not the least, Monasca API also offers faster query response times and paging. Its found to"},{"line_number":44,"context_line":"be much faster than the ceilometer v2 API [2]. It offers an interesting option to consider integrating"},{"line_number":45,"context_line":"Monasca API as a ceilometer storage driver or as an alternative for the ceilometer v2 API."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_97be03fe","line":45,"in_reply_to":"9a8ffd7b_facad94e","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"Integrate Monasca as publisher in ceilometer. A new publisher plugin needs to be implemented"},{"line_number":52,"context_line":"and will be enabled by entry_points and configured for metadata filtering, batching and"},{"line_number":53,"context_line":"failure retrying via conf."},{"line_number":54,"context_line":""},{"line_number":55,"context_line":"Metadata filtering is an important configuration for the publisher. It maps the resource metadata"},{"line_number":56,"context_line":"attributes for a meter that need to be preserved when the meter datapoint(sample) is published to"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_dad83520","line":53,"updated":"2015-11-24 21:16:10.000000000","message":"Ok, but you can already do that in Monasca, that\u0027s the point of the entry points system.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":57,"context_line":"Monasca. Metadata attributes that do not match in an incoming sample are discarded."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"The publisher plugin approach is preferred since it keep changes isolated from the main ceilometer tree"},{"line_number":60,"context_line":"and avoids the need of datapoints being re-published to RMQ after pipeline processing."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Pipeline processing will not be affected by the publisher since Monasca-api is async with respect"},{"line_number":63,"context_line":"to data persistence in backend."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_5a11a5ab","line":60,"updated":"2015-11-24 21:16:10.000000000","message":"That sounds good. I just wonder if you\u0027re going to publish to Kafka or HTTP Monasca API?","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":57,"context_line":"Monasca. Metadata attributes that do not match in an incoming sample are discarded."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"The publisher plugin approach is preferred since it keep changes isolated from the main ceilometer tree"},{"line_number":60,"context_line":"and avoids the need of datapoints being re-published to RMQ after pipeline processing."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Pipeline processing will not be affected by the publisher since Monasca-api is async with respect"},{"line_number":63,"context_line":"to data persistence in backend."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_375dd7e9","line":60,"in_reply_to":"9a8ffd7b_5a11a5ab","updated":"2015-11-24 22:59:48.000000000","message":"HTTP Monasca API","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":1669,"name":"Julien Danjou","display_name":"jd","email":"julien@danjou.info","username":"jdanjou"},"change_message_id":"c715712949d870eed90104b8efcc52a378600e61","unresolved":false,"context_lines":[{"line_number":121,"context_line":"* batch_mode, batch_count, batch_timeout, retry_interval, max_retries are some"},{"line_number":122,"context_line":"  of the new configuration params that will be defined in a new section - monasca"},{"line_number":123,"context_line":"  in ceilometer.conf. These however will have reasonable defaults."},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"* publisher will already be enabled via setup entry_points, but not be used by default."},{"line_number":126,"context_line":"  pipeline.yaml will need to include this publisher."},{"line_number":127,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_ba4e91bf","line":124,"updated":"2015-11-24 21:16:10.000000000","message":"It\u0027s not clear to me what the batch approach is proposing here. That sounds like you want to buffer things into the publisher part, which is a trade-off with respect to data integrity.","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":121,"context_line":"* batch_mode, batch_count, batch_timeout, retry_interval, max_retries are some"},{"line_number":122,"context_line":"  of the new configuration params that will be defined in a new section - monasca"},{"line_number":123,"context_line":"  in ceilometer.conf. These however will have reasonable defaults."},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"* publisher will already be enabled via setup entry_points, but not be used by default."},{"line_number":126,"context_line":"  pipeline.yaml will need to include this publisher."},{"line_number":127,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_971a8316","line":124,"in_reply_to":"9a8ffd7b_ba4e91bf","updated":"2015-11-24 22:59:48.000000000","message":"Yes, the idea is to buffer things in the publisher, but we can look at the other approaches to make this more consistent and fail-safe. At present, the notifier publisher also does some in-memory caching. [1]\n\n[1] https://github.com/openstack/ceilometer/blob/master/ceilometer/publisher/messaging.py#L136","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":12512,"name":"Deklan Dieterly","email":"deklan.dieterly@hpe.com","username":"deklan"},"change_message_id":"3ee98b54cae2ed3b503d7fb626916d271d6607e6","unresolved":false,"context_lines":[{"line_number":155,"context_line":"----------"},{"line_number":156,"context_line":""},{"line_number":157,"context_line":"* Add a new client facade for monasca client in ceilometer - monasca_client.py"},{"line_number":158,"context_line":"  to massage the calls made to monasca python bindings and handle common exceptions"},{"line_number":159,"context_line":""},{"line_number":160,"context_line":"* Create a new configuration file - monasca_field_definitions.yaml under etc/ceilometer to"},{"line_number":161,"context_line":"  enlist the metadata mappings for all ceilometer supported meters."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_df5fc70c","line":158,"updated":"2015-11-24 20:43:49.000000000","message":"period","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"8aca45cb025168624d546bc3df6abbc5492e54e6","unresolved":false,"context_lines":[{"line_number":155,"context_line":"----------"},{"line_number":156,"context_line":""},{"line_number":157,"context_line":"* Add a new client facade for monasca client in ceilometer - monasca_client.py"},{"line_number":158,"context_line":"  to massage the calls made to monasca python bindings and handle common exceptions"},{"line_number":159,"context_line":""},{"line_number":160,"context_line":"* Create a new configuration file - monasca_field_definitions.yaml under etc/ceilometer to"},{"line_number":161,"context_line":"  enlist the metadata mappings for all ceilometer supported meters."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9a8ffd7b_b71787ec","line":158,"in_reply_to":"9a8ffd7b_df5fc70c","updated":"2015-11-24 22:59:48.000000000","message":"Done","commit_id":"35800581409b2e8d4167dd593a75d9e2c25c74bf"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"8c2a060a394f3dd24fbb31c92bc5ac3e60393e77","unresolved":false,"context_lines":[{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The publisher can be enabled by including monasca:// as prefix for the publisher in a pipeline sink"},{"line_number":68,"context_line":"in the pipeline configuration file (pipeline.yaml)"},{"line_number":69,"context_line":""},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"Alternatives"},{"line_number":72,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"9a8ffd7b_ad7f0f65","line":69,"updated":"2015-11-25 15:28:11.000000000","message":"i think i mentioned this, but the in-memory batching is probably not ideal as you risk losing a great deal of information.\n\nyou should look into using the batching support in oslo.messaging.","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"e93fc21ea16db8754ef22d9742d73e8f92763a61","unresolved":false,"context_lines":[{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The publisher can be enabled by including monasca:// as prefix for the publisher in a pipeline sink"},{"line_number":68,"context_line":"in the pipeline configuration file (pipeline.yaml)"},{"line_number":69,"context_line":""},{"line_number":70,"context_line":""},{"line_number":71,"context_line":"Alternatives"},{"line_number":72,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"9a8ffd7b_9d2347a7","line":69,"in_reply_to":"9a8ffd7b_ad7f0f65","updated":"2015-12-01 23:37:59.000000000","message":"@gordc: The in-memory batching is a configurable option and can always be turned off, also the notifier (default) publisher also does some in-memory buffering [1].\n\nI think the batching support in oslo.messaging is good when consuming from rabbit, but in this case, Monasca api publishes to kafka and bypasses re-publishing to rabbit altogether.\n\n[1] https://github.com/openstack/ceilometer/blob/master/ceilometer/publisher/messaging.py#L136","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"8c2a060a394f3dd24fbb31c92bc5ac3e60393e77","unresolved":false,"context_lines":[{"line_number":76,"context_line":"This also uses a plugin approach, but it requires samples to be published back to RMQ"},{"line_number":77,"context_line":"(ceilometer metering bus) and will not improve write performance. While it does make the overall"},{"line_number":78,"context_line":"design more conformant to current integrations of different backends in ceilometer, the impact"},{"line_number":79,"context_line":"it has on load performance does make it unattractive to consider."},{"line_number":80,"context_line":""},{"line_number":81,"context_line":"Data model impact"},{"line_number":82,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"9a8ffd7b_8d41eb0d","line":79,"updated":"2015-11-25 15:28:11.000000000","message":"i think i also mentioned this in our email exchange but remind me why we can\u0027t do something like the graphite publisher[1] and maintain it in own repo? this would allow HP to control it as it pleases?\n\nif it\u0027s a matter of \u0027official\u0027 support, maybe it make sense to to move the repo under the telemetry project like what we did with powervm plugin\n\n[1] https://github.com/spilgames/ceilometer_graphite_publisher","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"e93fc21ea16db8754ef22d9742d73e8f92763a61","unresolved":false,"context_lines":[{"line_number":76,"context_line":"This also uses a plugin approach, but it requires samples to be published back to RMQ"},{"line_number":77,"context_line":"(ceilometer metering bus) and will not improve write performance. While it does make the overall"},{"line_number":78,"context_line":"design more conformant to current integrations of different backends in ceilometer, the impact"},{"line_number":79,"context_line":"it has on load performance does make it unattractive to consider."},{"line_number":80,"context_line":""},{"line_number":81,"context_line":"Data model impact"},{"line_number":82,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"9a8ffd7b_9df6e7f7","line":79,"in_reply_to":"9a8ffd7b_8d41eb0d","updated":"2015-12-01 23:37:59.000000000","message":"@gordc: Integrating monasca publisher in Ceilometer facilitates CI/CD and allows for integration issues to be identified faster. We also want more community involvement in it rather than just HP.  :)","commit_id":"b7998307fa8ed039ecf6dc20c9482dfc1be6030e"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":19,"context_line":"Ceilometer monitors and collects data from OpenStack via notifications sent"},{"line_number":20,"context_line":"from existing services or by polling the infrastructure."},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"When consuming large scale data, Ceilometer encounters bottleneck while republishing"},{"line_number":23,"context_line":"of samples to rabbitMQ. In large deployments, rabbitMQ clustering has some limitations"},{"line_number":24,"context_line":"around the 20M-mark load [1][2]; this can slow down the queue performance and impact other"},{"line_number":25,"context_line":"services relying on the queue."},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"We did extensive performance benchmarking of Ceilometer w/mongo and Ceilometer w/Monasca on"},{"line_number":28,"context_line":"VM and baremetal boxes and found Ceilometer w/monasca to be 3x faster than Ceilometer to"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_fb2a968f","line":25,"range":{"start_line":22,"start_character":0,"end_line":25,"end_character":30},"updated":"2015-12-02 03:45:54.000000000","message":"even with multiple rabbit host? this is probably advised regardless as you don\u0027t want different services interfering with others.","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":22,"context_line":"When consuming large scale data, Ceilometer encounters bottleneck while republishing"},{"line_number":23,"context_line":"of samples to rabbitMQ. In large deployments, rabbitMQ clustering has some limitations"},{"line_number":24,"context_line":"around the 20M-mark load [1][2]; this can slow down the queue performance and impact other"},{"line_number":25,"context_line":"services relying on the queue."},{"line_number":26,"context_line":""},{"line_number":27,"context_line":"We did extensive performance benchmarking of Ceilometer w/mongo and Ceilometer w/Monasca on"},{"line_number":28,"context_line":"VM and baremetal boxes and found Ceilometer w/monasca to be 3x faster than Ceilometer to"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_7e268fbb","line":25,"in_reply_to":"9a8ffd7b_fb2a968f","updated":"2015-12-02 07:13:32.000000000","message":"This was encountered with single rabbit host which come with the devstack deployment, but more than impacting other services, rabbit itself faced issues keeping up at load\u003e15M","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":30,"context_line":"[4] that we tested with exists in a separate repo."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"Batching of samples is an important feature to achieve high insert throughput. Most ceilometer"},{"line_number":33,"context_line":"backends do not support batching and the ones that do are for datapoints for the same meter,"},{"line_number":34,"context_line":"whereas in a cloud, samples arrive in random for disparate meters. Its important for the backend"},{"line_number":35,"context_line":"to support batching for cross-meter samples."},{"line_number":36,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_fb013607","line":33,"updated":"2015-12-02 03:45:54.000000000","message":"i\u0027m not sure what this statement refers to? which backend is batching on the same meter?","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":30,"context_line":"[4] that we tested with exists in a separate repo."},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"Batching of samples is an important feature to achieve high insert throughput. Most ceilometer"},{"line_number":33,"context_line":"backends do not support batching and the ones that do are for datapoints for the same meter,"},{"line_number":34,"context_line":"whereas in a cloud, samples arrive in random for disparate meters. Its important for the backend"},{"line_number":35,"context_line":"to support batching for cross-meter samples."},{"line_number":36,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_7e2befeb","line":33,"in_reply_to":"9a8ffd7b_fb013607","updated":"2015-12-02 07:13:32.000000000","message":"The gnocchi API to post a list of measures for a metric: POST /v1/entity/\u003cname\u003e/measures (https://wiki.openstack.org/wiki/Gnocchi) in contrast to the monasca API to post an array of metrics (https://github.com/openstack/monasca-api/blob/master/docs/monasca-api-spec.md#array-of-metrics)","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":61,"context_line":"Ceilometer has a default publisher (notifier://) that buffers samples in-memory before"},{"line_number":62,"context_line":"publishing a batch.[5]"},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"Metadata filtering is an important configuration for the publisher. It maps the resource metadata"},{"line_number":65,"context_line":"attributes for a meter that need to be preserved when the meter datapoint(sample) is published to"},{"line_number":66,"context_line":"Monasca. Metadata attributes that do not match in an incoming sample are discarded."},{"line_number":67,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_d64be957","line":64,"updated":"2015-12-02 03:45:54.000000000","message":"this filtering is separate from pipeline?","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":61,"context_line":"Ceilometer has a default publisher (notifier://) that buffers samples in-memory before"},{"line_number":62,"context_line":"publishing a batch.[5]"},{"line_number":63,"context_line":""},{"line_number":64,"context_line":"Metadata filtering is an important configuration for the publisher. It maps the resource metadata"},{"line_number":65,"context_line":"attributes for a meter that need to be preserved when the meter datapoint(sample) is published to"},{"line_number":66,"context_line":"Monasca. Metadata attributes that do not match in an incoming sample are discarded."},{"line_number":67,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_1eabc3d7","line":64,"in_reply_to":"9a8ffd7b_d64be957","updated":"2015-12-02 07:13:32.000000000","message":"Yes, this filtering is done in the publisher after the pipeline processing","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":81,"context_line":"Integrate Monasca as a ceilometer dispatcher."},{"line_number":82,"context_line":""},{"line_number":83,"context_line":"This also uses a plugin approach and can leverage batching at the oslo.messaging level[6] to consume"},{"line_number":84,"context_line":"samples in a batch, but it requires samples to be published back to RMQ (ceilometer metering bus). which"},{"line_number":85,"context_line":"is redundant in this case, since Monasca has its own transport that it leverages when posting measurements."},{"line_number":86,"context_line":"We also found rabbit to have scaling issues at higher loads whereas Kafka scaled better."},{"line_number":87,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_d66209d8","line":84,"updated":"2015-12-02 03:45:54.000000000","message":"these last two paragraphs aren\u0027t actually true. see: https://review.openstack.org/#/c/250814/","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":81,"context_line":"Integrate Monasca as a ceilometer dispatcher."},{"line_number":82,"context_line":""},{"line_number":83,"context_line":"This also uses a plugin approach and can leverage batching at the oslo.messaging level[6] to consume"},{"line_number":84,"context_line":"samples in a batch, but it requires samples to be published back to RMQ (ceilometer metering bus). which"},{"line_number":85,"context_line":"is redundant in this case, since Monasca has its own transport that it leverages when posting measurements."},{"line_number":86,"context_line":"We also found rabbit to have scaling issues at higher loads whereas Kafka scaled better."},{"line_number":87,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_dee0db98","line":84,"in_reply_to":"9a8ffd7b_d66209d8","updated":"2015-12-02 07:13:32.000000000","message":"Not sure whats incorrect here, i meant having a dispatcher means using the ceilometer metering bus and the batching support in the messaging layer should help, but its probably good if rabbit is the only choice as transport and could lead to speedups, but in this case, its not needed since Monasca API publishes to its internal kafka transport. Let me know if i missed anything.","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":107,"context_line":"---------------"},{"line_number":108,"context_line":""},{"line_number":109,"context_line":"Monasca is a multi-tenant solution so tenant-level segregation of datapoints"},{"line_number":110,"context_line":"is enforced at all times."},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"Pipeline impact"},{"line_number":113,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_d61d2967","line":110,"updated":"2015-12-02 03:45:54.000000000","message":"does kafka have any security support yet?","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":107,"context_line":"---------------"},{"line_number":108,"context_line":""},{"line_number":109,"context_line":"Monasca is a multi-tenant solution so tenant-level segregation of datapoints"},{"line_number":110,"context_line":"is enforced at all times."},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"Pipeline impact"},{"line_number":113,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_befcf73e","line":110,"in_reply_to":"9a8ffd7b_d61d2967","updated":"2015-12-02 07:13:32.000000000","message":"Not sure on this, will check and get back.","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":123,"context_line":"Performance/Scalability Impacts"},{"line_number":124,"context_line":"-------------------------------"},{"line_number":125,"context_line":""},{"line_number":126,"context_line":"* Monasca publisher can support much higher rates of message consumption."},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"* Data retention policies will be supported and enforced by Monasca."},{"line_number":129,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_96824151","line":126,"updated":"2015-12-02 03:45:54.000000000","message":"what does publisher have to do with consumption? i\u0027d probably remove this statement","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":123,"context_line":"Performance/Scalability Impacts"},{"line_number":124,"context_line":"-------------------------------"},{"line_number":125,"context_line":""},{"line_number":126,"context_line":"* Monasca publisher can support much higher rates of message consumption."},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"* Data retention policies will be supported and enforced by Monasca."},{"line_number":129,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_3eca47a2","line":126,"in_reply_to":"9a8ffd7b_96824151","updated":"2015-12-02 07:13:32.000000000","message":"I meant Monasca API and Monasca in general.","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":125,"context_line":""},{"line_number":126,"context_line":"* Monasca publisher can support much higher rates of message consumption."},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"* Data retention policies will be supported and enforced by Monasca."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"Other deployer impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_f68b8d66","line":128,"updated":"2015-12-02 03:45:54.000000000","message":"probably not relevant to publisher","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":125,"context_line":""},{"line_number":126,"context_line":"* Monasca publisher can support much higher rates of message consumption."},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"* Data retention policies will be supported and enforced by Monasca."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":""},{"line_number":131,"context_line":"Other deployer impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_fec4bf92","line":128,"in_reply_to":"9a8ffd7b_f68b8d66","updated":"2015-12-02 07:13:32.000000000","message":"Done","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":6537,"name":"gordon chung","email":"gord@live.ca","username":"chungg"},"change_message_id":"f99a2015f2831144990bcf5d250dabde80d43cee","unresolved":false,"context_lines":[{"line_number":140,"context_line":""},{"line_number":141,"context_line":"* new configuration file for metadata mappings will need to be copied over to /etc/ceilometer"},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"* This depends on python-monascaclient, so that package will need to be included in requirements.txt"},{"line_number":144,"context_line":""},{"line_number":145,"context_line":""},{"line_number":146,"context_line":"Developer impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_f6e2ad2e","line":143,"updated":"2015-12-02 03:45:54.000000000","message":"test-requirements.txt. same as driver clients (mongo, gnocchi, etc...)","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"},{"author":{"_account_id":13560,"name":"Rohit Jaiswal","email":"rohit.jaiswal@gmail.com","username":"rjaiswal"},"change_message_id":"1aa19f5197bb199f9f24c0878e1e89ad33b5e07b","unresolved":false,"context_lines":[{"line_number":140,"context_line":""},{"line_number":141,"context_line":"* new configuration file for metadata mappings will need to be copied over to /etc/ceilometer"},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"* This depends on python-monascaclient, so that package will need to be included in requirements.txt"},{"line_number":144,"context_line":""},{"line_number":145,"context_line":""},{"line_number":146,"context_line":"Developer impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9a8ffd7b_1ec0a37b","line":143,"in_reply_to":"9a8ffd7b_f6e2ad2e","updated":"2015-12-02 07:13:32.000000000","message":"Done","commit_id":"db7f415234d64cba081a2622883c76abf38dc691"}]}