This test run is using yoga/jammy. 39 tempest tests failed, and they are mostly related to timeouts reaching neutron. The load figures show that memory runs very low towards the end and the neutron leader machine (server-10) starts using the swap memory. I suppose that could lead to very slow responses.
Please let me know what you think. Also, if you have different metrics you wish to see we can update the automation to collect more.
We have some load figures, for testrun eff35558- 063b-4313- 8ee6-30a406a878 2d:
Build output (jujus status at the bottom): https:/ /oil-jenkins. canonical. com/job/ fce_build/ 27515// console /solutions. qa.canonical. com/grafana/ d/r73je3CVz/ test-run- metrics? orgId=1& var-testrun_ uuid=eff35558- 063b-4313- 8ee6-30a406a878 2d&var- metric= mem_free& var-juju_ model=openstack &from=169171308 8000&to= 1691727038000
Load figures: https:/
This test run is using yoga/jammy. 39 tempest tests failed, and they are mostly related to timeouts reaching neutron. The load figures show that memory runs very low towards the end and the neutron leader machine (server-10) starts using the swap memory. I suppose that could lead to very slow responses.
Please let me know what you think. Also, if you have different metrics you wish to see we can update the automation to collect more.