On Thu, Dec 6, 2018 at 9:26 PM Whaley, Graham <graham.whaley@intel.com> wrote:
-----Original Message----- From: Whaley, Graham [mailto:graham.whaley@intel.com] Sent: Thursday, December 6, 2018 9:40 AM To: kata-dev <kata-dev@lists.katacontainers.io> Subject: [kata-dev] metrics CI failing at present
Hi all, The metrics CI has been nack'ing PRs for the last 48h or so - it feels like maybe some PRs have landed that have effected boot time and memory footprint. I'll go track down when that happened, send an update here, and adjust the metrics bounds check values if needed. In the meantime, don't panic about metrics CI failures...
Graham
Looks like a broken PR build/run left a bunch of containerd-shims around that threw out the metrics and were not cleaned up by the bare metal cleanup scripts: https://github.com/kata-containers/ci/issues/94
A reboot of the slave seems to have gotten the metrics builds back to passing. http://jenkins.katacontainers.io/computer/x86_packet01/builds
Let's assume 'normal service resumed', and report any more metrics fails we see.
Hi Graham, It seems metrics is still failing, see http://jenkins.katacontainers.io/job/kata-metrics-agent-ubuntu-16-04-PR/ Cheers, Tao -- bergwolf@hyper.sh