Description of problem: The issue with extracting reliable usage and event data from openstack with gnocchi. The gnocchi resource "started_at" timestamps are not the correct timestamps of the nova instance start time, as they differ a few minutes from the nova data and the ceilometer event data for the instance start event. The started_at is set the first time that Ceilometer creates the resource in Gnocchi, no data from Nova is used. Gnocchi resources never get an “ended_at” date. Version-Release number of selected component (if applicable): Gnocchi/Ceilometer in OSP10 How reproducible: The gnocchi resource of type instance has the following timestamp: (gnocchi resource list or gnocchi resource show) Resource started at: 2017-12-01T10:11:34.724926+00:00 Revision started at: 2017-12-01T10:11:34.724948+00:00 nova show delivers the following for the launched_at param of nova: OS-SRV-USG:launched_at | 2017-12-01T10:07:36.000000 When looking up the ceilometer compute.instance.create.end event, this has the timestamp: ceilometer event-list -q 'event_type=compute.instance.create.end;instance_id=5023b71f-c8c2-4a1a-82f0-1fdafdb090b7' launched_at | datetime | 2017-12-01T10:07:36.759000 The according ceilometer event was generated at: 2017-12-01T10:07:37.017000 Actual results: Gnocchi started_at timestamps are not in sync with the timestamps of the ceilometer events. There are nearly 4 minutes in between the launched_at timestamp of the instance and the "started_at" timestamp of gnocchi. Gnocchi ended_at timestamps are missing. There is also (small) difference in between the launched_at time value in nova and the ceilometer event. Expected results: Customer needs reliable data for metering and billing. At the moment It looks like every component creates an own timestamp.
Tested, automated. Verified
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2019:0045