Add missing-container-metrics
node-exporter
exposes the node_vmstat_oom_kill
metric that increases when the host reports an OOM kill. However, that doesn't tell us which pod/container the killed process was a part of.
There are some projects that aim to fill that gap. We have used missing-container-metrics
on stackspin.net
for a few weeks, and it seems to do the job. Also we're having OOMs with some frequency, so it would be really useful to be able to have more information.