)]}'
{"doc/source/specs/prometheus-exporter.rst":[{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"ff1084a548cb767eff52457d53a66299c3791746","unresolved":false,"context_lines":[{"line_number":39,"context_line":"it under OpenStack-Helm umbrella. Basic idea is to combine both API polling and"},{"line_number":40,"context_line":"notification listening to achieve the goal of exposing the metrics."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"On the start the exporter will get current state of the system using OpenStack"},{"line_number":43,"context_line":"API queries while starting to listen to notifications from OpenStack services."},{"line_number":44,"context_line":"It\u0027s important to say that those notifications will need to be turned on in"},{"line_number":45,"context_line":"each OpenStack service (otherwise the exporter will be similar to Canonical\u0027s"},{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conviniently saved into a Redis"}],"source_content_type":"text/x-rst","patch_set":1,"id":"df3967d1_df3c6be2","line":46,"range":{"start_line":42,"start_character":0,"end_line":46,"end_character":4},"updated":"2017-08-04 13:29:58.000000000","message":"One approach to consider is reusing some of the code used for the Fuel LMA plugin: https://github.com/openstack/fuel-plugin-lma-collector/tree/master/deployment_scripts/puppet/modules/lma_collector/files/collectd.\n\nUsing some of the code here could be valuable as it gathers metrics we\u0027d like to have, and also goes beyond openstack to include ceph as well. This could reduce the overhead associated with writing something from scratch.","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":11600,"name":"Michał Dulko","email":"michal.dulko@gmail.com","username":"dulek"},"change_message_id":"3827fc8dea21326397370d4957e45c5553386fa8","unresolved":false,"context_lines":[{"line_number":39,"context_line":"it under OpenStack-Helm umbrella. Basic idea is to combine both API polling and"},{"line_number":40,"context_line":"notification listening to achieve the goal of exposing the metrics."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"On the start the exporter will get current state of the system using OpenStack"},{"line_number":43,"context_line":"API queries while starting to listen to notifications from OpenStack services."},{"line_number":44,"context_line":"It\u0027s important to say that those notifications will need to be turned on in"},{"line_number":45,"context_line":"each OpenStack service (otherwise the exporter will be similar to Canonical\u0027s"},{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conviniently saved into a Redis"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_082ed242","line":46,"range":{"start_line":42,"start_character":0,"end_line":46,"end_character":4},"in_reply_to":"df3967d1_df3c6be2","updated":"2017-08-09 09:10:27.000000000","message":"Oh, that\u0027s pretty useful, thanks.","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"ff1084a548cb767eff52457d53a66299c3791746","unresolved":false,"context_lines":[{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conviniently saved into a Redis"},{"line_number":50,"context_line":"database. Please note that using simple, non-persistent, NoSQL DB is fine - the"},{"line_number":51,"context_line":"state can always be recreated by doing an OpenStack API query."},{"line_number":52,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"df3967d1_22059c11","line":49,"range":{"start_line":49,"start_character":65,"end_line":49,"end_character":70},"updated":"2017-08-04 13:29:58.000000000","message":"Should we add a Redis chart to the work items?","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":11600,"name":"Michał Dulko","email":"michal.dulko@gmail.com","username":"dulek"},"change_message_id":"3827fc8dea21326397370d4957e45c5553386fa8","unresolved":false,"context_lines":[{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conviniently saved into a Redis"},{"line_number":50,"context_line":"database. Please note that using simple, non-persistent, NoSQL DB is fine - the"},{"line_number":51,"context_line":"state can always be recreated by doing an OpenStack API query."},{"line_number":52,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_08d2525d","line":49,"range":{"start_line":49,"start_character":65,"end_line":49,"end_character":70},"in_reply_to":"df3967d1_22059c11","updated":"2017-08-09 09:10:27.000000000","message":"Whoops, good catch. I\u0027ve initially thought of standalone SQLite DB, but then realized that Redis is more free-form so it could fit better. I haven\u0027t thought that it would be a bad practice to put it inside the container with the exporter itself.\n\nI think I can switch that to etcd, as we\u0027re deploying it already.","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"ff1084a548cb767eff52457d53a66299c3791746","unresolved":false,"context_lines":[{"line_number":60,"context_line":"Metrics"},{"line_number":61,"context_line":"-------"},{"line_number":62,"context_line":"Intially set of measured metrics will be the same as of"},{"line_number":63,"context_line":"prometheus-openstack-exporter for Nova and Neutron. Swift is excluded here,"},{"line_number":64,"context_line":"because OpenStack-Helm is still missing a chart for Swift."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Nova"},{"line_number":67,"context_line":"~~~~"}],"source_content_type":"text/x-rst","patch_set":1,"id":"df3967d1_a2acecec","line":64,"range":{"start_line":63,"start_character":51,"end_line":64,"end_character":58},"updated":"2017-08-04 13:29:58.000000000","message":"Swift is on the roadmap.  We should include the framework for including Swift, but make it a noop until the chart is completed","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":11600,"name":"Michał Dulko","email":"michal.dulko@gmail.com","username":"dulek"},"change_message_id":"3827fc8dea21326397370d4957e45c5553386fa8","unresolved":false,"context_lines":[{"line_number":60,"context_line":"Metrics"},{"line_number":61,"context_line":"-------"},{"line_number":62,"context_line":"Intially set of measured metrics will be the same as of"},{"line_number":63,"context_line":"prometheus-openstack-exporter for Nova and Neutron. Swift is excluded here,"},{"line_number":64,"context_line":"because OpenStack-Helm is still missing a chart for Swift."},{"line_number":65,"context_line":""},{"line_number":66,"context_line":"Nova"},{"line_number":67,"context_line":"~~~~"}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_0876d2a5","line":64,"range":{"start_line":63,"start_character":51,"end_line":64,"end_character":58},"in_reply_to":"df3967d1_a2acecec","updated":"2017-08-09 09:10:27.000000000","message":"Sure thing, it should be highly plugable.","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"ff1084a548cb767eff52457d53a66299c3791746","unresolved":false,"context_lines":[{"line_number":124,"context_line":"Primary assignee:"},{"line_number":125,"context_line":"  mateusz-blaszkowski-8 (mateuszb)"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"Work Items"},{"line_number":128,"context_line":"----------"},{"line_number":129,"context_line":"- Decide on additional metrics we want to gather."},{"line_number":130,"context_line":"- Write implementation of API querying mechanism."}],"source_content_type":"text/x-rst","patch_set":1,"id":"df3967d1_42d54043","line":127,"range":{"start_line":127,"start_character":0,"end_line":127,"end_character":10},"updated":"2017-08-04 13:29:58.000000000","message":"It\u0027s outside the scope of the exporter, but we should include a work item for adding the appropriate configurations into Prometheus to scrape the metrics being exported, as it\u0027s required for the Prometheus tests mentioned","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":11600,"name":"Michał Dulko","email":"michal.dulko@gmail.com","username":"dulek"},"change_message_id":"3827fc8dea21326397370d4957e45c5553386fa8","unresolved":false,"context_lines":[{"line_number":124,"context_line":"Primary assignee:"},{"line_number":125,"context_line":"  mateusz-blaszkowski-8 (mateuszb)"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":"Work Items"},{"line_number":128,"context_line":"----------"},{"line_number":129,"context_line":"- Decide on additional metrics we want to gather."},{"line_number":130,"context_line":"- Write implementation of API querying mechanism."}],"source_content_type":"text/x-rst","patch_set":1,"id":"9f436f4f_68667ef0","line":127,"range":{"start_line":127,"start_character":0,"end_line":127,"end_character":10},"in_reply_to":"df3967d1_42d54043","updated":"2017-08-09 09:10:27.000000000","message":"Good point.","commit_id":"03bf2c73f3026fba9d5b75bf21c8e5e174c9cdf7"},{"author":{"_account_id":18503,"name":"Rakesh Patnaik","email":"patsrakesh@gmail.com","username":"rp196m"},"change_message_id":"97fc8add2b3e4e5b60723c55c5d84a2aef5c617d","unresolved":false,"context_lines":[{"line_number":82,"context_line":"* ``nova_resources_ram_mbs`` (per cloud, tenant)"},{"line_number":83,"context_line":"* ``nova_resources_vcpus`` (per cloud, tenant)"},{"line_number":84,"context_line":"* ``nova_resources_disk_gbs`` (per cloud, tenant)"},{"line_number":85,"context_line":""},{"line_number":86,"context_line":"Neutron"},{"line_number":87,"context_line":"~~~~~~~"},{"line_number":88,"context_line":"* ``neutron_net_size`` (per cloud and network)"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9f436f4f_4c9900c2","line":85,"updated":"2017-08-15 07:03:53.000000000","message":"nova control pane services and their availability is required for operational monitoring of the under cloud. We could use v2/os-services call https://developer.openstack.org/api-ref/compute/#list-compute-services to pull a state of the services and their state per controller, compute node etc.\n\nhost\u003dcontroller1,service\u003dnova-scheduler,state\u003dup\nhost\u003dcontroller2,service\u003dnova-scheduler,state\u003ddown\nhost\u003dcompute1,service\u003dnova-compute,state\u003dup","commit_id":"f73bf03a11aca05a27f498f18ecac1655bfe5a7f"},{"author":{"_account_id":11636,"name":"Mateusz Blaszkowski","email":"mateusz.blaszkowski@intel.com","username":"mateuszb"},"change_message_id":"662539c69312622c54e869615aff51475d09516c","unresolved":false,"context_lines":[{"line_number":82,"context_line":"* ``nova_resources_ram_mbs`` (per cloud, tenant)"},{"line_number":83,"context_line":"* ``nova_resources_vcpus`` (per cloud, tenant)"},{"line_number":84,"context_line":"* ``nova_resources_disk_gbs`` (per cloud, tenant)"},{"line_number":85,"context_line":""},{"line_number":86,"context_line":"Neutron"},{"line_number":87,"context_line":"~~~~~~~"},{"line_number":88,"context_line":"* ``neutron_net_size`` (per cloud and network)"}],"source_content_type":"text/x-rst","patch_set":2,"id":"9f436f4f_3a010f6b","line":85,"in_reply_to":"9f436f4f_4c9900c2","updated":"2017-08-17 06:26:30.000000000","message":"Since the metrics we collect may be presented on Prometheus graph, I\u0027d recommend to gather numeric values like nova_computes_running and nova_computes_up (the same for scheduler, etc). This will also allow us to easily add alerts to Prometheus (like less than X percent of API instances running).","commit_id":"f73bf03a11aca05a27f498f18ecac1655bfe5a7f"},{"author":{"_account_id":18503,"name":"Rakesh Patnaik","email":"patsrakesh@gmail.com","username":"rp196m"},"change_message_id":"97fc8add2b3e4e5b60723c55c5d84a2aef5c617d","unresolved":false,"context_lines":[{"line_number":87,"context_line":"~~~~~~~"},{"line_number":88,"context_line":"* ``neutron_net_size`` (per cloud and network)"},{"line_number":89,"context_line":"* ``neutron_public_ip_usage`` (per cloud, subnet, tenant, ip type, ip status)"},{"line_number":90,"context_line":""},{"line_number":91,"context_line":"Other services"},{"line_number":92,"context_line":"~~~~~~~~~~~~~~"},{"line_number":93,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9f436f4f_4c758063","line":90,"updated":"2017-08-15 07:03:53.000000000","message":"A frequent issue in operations is to maintain neutron agents and their state. we could use v2.0/agents to pull the state of agents. https://specs.openstack.org/openstack/neutron-specs/specs/api/agent_management.html#list-agents\n\nhost\u003drouter1,service\u003dneutron-dhcp-agent,state\u003dup\nhost\u003drouter2,service\u003dneutron-lbaas-agent,state\u003ddown","commit_id":"f73bf03a11aca05a27f498f18ecac1655bfe5a7f"},{"author":{"_account_id":11636,"name":"Mateusz Blaszkowski","email":"mateusz.blaszkowski@intel.com","username":"mateuszb"},"change_message_id":"662539c69312622c54e869615aff51475d09516c","unresolved":false,"context_lines":[{"line_number":87,"context_line":"~~~~~~~"},{"line_number":88,"context_line":"* ``neutron_net_size`` (per cloud and network)"},{"line_number":89,"context_line":"* ``neutron_public_ip_usage`` (per cloud, subnet, tenant, ip type, ip status)"},{"line_number":90,"context_line":""},{"line_number":91,"context_line":"Other services"},{"line_number":92,"context_line":"~~~~~~~~~~~~~~"},{"line_number":93,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9f436f4f_1afc0b6e","line":90,"in_reply_to":"9f436f4f_4c758063","updated":"2017-08-17 06:26:30.000000000","message":"The same applies to Neutron services - I believe gathering the sum of up and total services will be more adequate.","commit_id":"f73bf03a11aca05a27f498f18ecac1655bfe5a7f"},{"author":{"_account_id":8181,"name":"Jaesuk Ahn","email":"bluejay.ahn@gmail.com","username":"blujeay.kr"},"change_message_id":"1ae1b1c624be080ddd574e6e1d35243099a35a84","unresolved":false,"context_lines":[{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conviniently saved into an etcd"},{"line_number":50,"context_line":"database. Please note that using simple, non-persistent, NoSQL DB is fine - the"},{"line_number":51,"context_line":"state can always be recreated by doing an OpenStack API query."},{"line_number":52,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"7f287b81_f92a528e","line":49,"range":{"start_line":49,"start_character":39,"end_line":49,"end_character":51},"updated":"2017-08-23 01:29:31.000000000","message":"typo: conviniently \u003e\u003e conveniently","commit_id":"defb91c105c4b56639da5eddc260d6227b2b878c"},{"author":{"_account_id":8181,"name":"Jaesuk Ahn","email":"bluejay.ahn@gmail.com","username":"blujeay.kr"},"change_message_id":"1ae1b1c624be080ddd574e6e1d35243099a35a84","unresolved":false,"context_lines":[{"line_number":59,"context_line":""},{"line_number":60,"context_line":"Metrics"},{"line_number":61,"context_line":"-------"},{"line_number":62,"context_line":"Intially set of measured metrics will be the same as of"},{"line_number":63,"context_line":"prometheus-openstack-exporter for Nova and Neutron. Swift is excluded here,"},{"line_number":64,"context_line":"because OpenStack-Helm is still missing a chart for Swift."},{"line_number":65,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"7f287b81_b91e0a6e","line":62,"range":{"start_line":62,"start_character":0,"end_line":62,"end_character":8},"updated":"2017-08-23 01:29:31.000000000","message":"typo: Intially \u003e\u003e Initially","commit_id":"defb91c105c4b56639da5eddc260d6227b2b878c"},{"author":{"_account_id":15518,"name":"Artur Korzeniewski","email":"artur.korzen@wp.pl","username":"akorzeni"},"change_message_id":"a5c31dc91e8c1876093b9e34f8fb0366ad39d1db","unresolved":false,"context_lines":[{"line_number":16,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":17,"context_line":"Currently our choice for metrics database and analysis engine is `Prometheus"},{"line_number":18,"context_line":"\u003chttps://prometheus.io/\u003e`_. While various apps like k8s or etcd have endpoints"},{"line_number":19,"context_line":"with metrics data in Prometheus format, OpenStack services aren\u0027t. Therefore to"},{"line_number":20,"context_line":"utilize Prometheus we should use external metric aggregator."},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"Current draft of the implementation uses `prometheus-openstack-exporter"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_1b6c4c06","line":19,"range":{"start_line":19,"start_character":59,"end_line":19,"end_character":65},"updated":"2017-10-02 05:42:48.000000000","message":"don\u0027t have?","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":14250,"name":"Grzegorz Grasza","email":"xek@redhat.com","username":"xek"},"change_message_id":"314f6e3df4b7a4f18f68e4e0fc078f8200832652","unresolved":false,"context_lines":[{"line_number":17,"context_line":"Currently our choice for metrics database and analysis engine is `Prometheus"},{"line_number":18,"context_line":"\u003chttps://prometheus.io/\u003e`_. While various apps like k8s or etcd have endpoints"},{"line_number":19,"context_line":"with metrics data in Prometheus format, OpenStack services aren\u0027t. Therefore to"},{"line_number":20,"context_line":"utilize Prometheus we should use external metric aggregator."},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"Current draft of the implementation uses `prometheus-openstack-exporter"},{"line_number":23,"context_line":"\u003chttps://github.com/CanonicalLtd/prometheus-openstack-exporter\u003e`_, but this"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_66949137","line":20,"range":{"start_line":20,"start_character":32,"end_line":20,"end_character":33},"updated":"2017-10-03 09:59:00.000000000","message":"an","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":15518,"name":"Artur Korzeniewski","email":"artur.korzen@wp.pl","username":"akorzeni"},"change_message_id":"a5c31dc91e8c1876093b9e34f8fb0366ad39d1db","unresolved":false,"context_lines":[{"line_number":21,"context_line":""},{"line_number":22,"context_line":"Current draft of the implementation uses `prometheus-openstack-exporter"},{"line_number":23,"context_line":"\u003chttps://github.com/CanonicalLtd/prometheus-openstack-exporter\u003e`_, but this"},{"line_number":24,"context_line":"blueprint is proposing writing our own solution to the problem."},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"Use cases"},{"line_number":27,"context_line":"---------"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_9b7f1caa","line":24,"range":{"start_line":24,"start_character":10,"end_line":24,"end_character":21},"updated":"2017-10-02 05:42:48.000000000","message":"proposes","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"a4188489172999019f77cab3057aff621cd14e6a","unresolved":false,"context_lines":[{"line_number":39,"context_line":"it under OpenStack-Helm umbrella. Basic idea is to combine both API polling and"},{"line_number":40,"context_line":"notification listening to achieve the goal of exposing the metrics."},{"line_number":41,"context_line":""},{"line_number":42,"context_line":"On the start the exporter will get current state of the system using OpenStack"},{"line_number":43,"context_line":"API queries while starting to listen to notifications from OpenStack services."},{"line_number":44,"context_line":"It\u0027s important to say that those notifications will need to be turned on in"},{"line_number":45,"context_line":"each OpenStack service (otherwise the exporter will be similar to Canonical\u0027s"},{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conveniently saved into an etcd"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_3f5f970f","line":46,"range":{"start_line":42,"start_character":0,"end_line":46,"end_character":4},"updated":"2017-10-04 17:51:35.000000000","message":"The metrics related to the openstack services and the notifications from those services should be separate. I dont think a solution should encompass both for our needs, as that\u0027s not the problem we\u0027re looking to solve here.","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"a4188489172999019f77cab3057aff621cd14e6a","unresolved":false,"context_lines":[{"line_number":45,"context_line":"each OpenStack service (otherwise the exporter will be similar to Canonical\u0027s"},{"line_number":46,"context_line":"one)."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conveniently saved into an etcd"},{"line_number":50,"context_line":"database. Please note that using simple, non-persistent, NoSQL DB is fine - the"},{"line_number":51,"context_line":"state can always be recreated by doing an OpenStack API query."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"OpenStack API queries will be executed periodically to make sure state produced"},{"line_number":54,"context_line":"by notification listening is valid and consistent. In case of inconsistency"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_7fcbcf71","line":51,"range":{"start_line":48,"start_character":0,"end_line":51,"end_character":62},"updated":"2017-10-04 17:51:35.000000000","message":"I\u0027m okay with caching data, similar to the Canonical approach with their exporter, but I don\u0027t think we need to back an exporter with etcd or a database, as it just introduces yet another dependency for what\u0027s supposed to be a lightweight service","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":14250,"name":"Grzegorz Grasza","email":"xek@redhat.com","username":"xek"},"change_message_id":"314f6e3df4b7a4f18f68e4e0fc078f8200832652","unresolved":false,"context_lines":[{"line_number":48,"context_line":"By monitoring notifications the exporter will be able to make changes to it\u0027s"},{"line_number":49,"context_line":"state (i.e. metrics) and state will be conveniently saved into an etcd"},{"line_number":50,"context_line":"database. Please note that using simple, non-persistent, NoSQL DB is fine - the"},{"line_number":51,"context_line":"state can always be recreated by doing an OpenStack API query."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"OpenStack API queries will be executed periodically to make sure state produced"},{"line_number":54,"context_line":"by notification listening is valid and consistent. In case of inconsistency"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_66bdb1b5","line":51,"updated":"2017-10-03 09:59:00.000000000","message":"Maybe a cache of these values will be sufficient then? Memcached could be simpler to maintain. How much data are we talking about?","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":15518,"name":"Artur Korzeniewski","email":"artur.korzen@wp.pl","username":"akorzeni"},"change_message_id":"a5c31dc91e8c1876093b9e34f8fb0366ad39d1db","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"OpenStack API queries will be executed periodically to make sure state produced"},{"line_number":54,"context_line":"by notification listening is valid and consistent. In case of inconsistency"},{"line_number":55,"context_line":"state will be invalidated and API value will be used."},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"Those metrics will be exposed to Prometheus on ``/metrics`` endpoint without"},{"line_number":58,"context_line":"any authorization."}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_7b53403a","line":55,"range":{"start_line":55,"start_character":33,"end_line":55,"end_character":39},"updated":"2017-10-02 05:42:48.000000000","message":"values or output","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"},{"author":{"_account_id":17591,"name":"Steve Wilkerson","email":"wilkers.steve@gmail.com","username":"srwilkers"},"change_message_id":"a4188489172999019f77cab3057aff621cd14e6a","unresolved":false,"context_lines":[{"line_number":123,"context_line":"quite a useful way, log gathering is based on costly API queries. A better"},{"line_number":124,"context_line":"alternative should be based on notifications."},{"line_number":125,"context_line":""},{"line_number":126,"context_line":"`OpenStack Ceilometer exporter for Prometheus"},{"line_number":127,"context_line":"\u003chttps://github.com/carlpett/openstack-ceilometer_exporter\u003e`_ can gather and"},{"line_number":128,"context_line":"expose any metrics provided by Ceilometer. Beside the fact that project was"},{"line_number":129,"context_line":"last time active in August 2016, problem with that idea is that you need"},{"line_number":130,"context_line":"Ceilometer with all of its components running to aquire metrics."},{"line_number":131,"context_line":""},{"line_number":132,"context_line":"Implementation"},{"line_number":133,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":7,"id":"7f515b1d_3fec37f2","line":130,"range":{"start_line":126,"start_character":0,"end_line":130,"end_character":64},"updated":"2017-10-04 17:51:35.000000000","message":"OpenStack-Helm now has charts for ceilometer and its dependencies, but I feel this would target metrics specific to tenant usage as opposed to metrics about the infrastructure services we\u0027re concerned about.","commit_id":"3a4a7c72b794e5b9cfa67cfaa5793468a1c122b1"}]}
