Create alert for OOMKill events inside containers #822
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
paulfantom
approved these changes
Jan 31, 2023
|
This is very similiar to #800 see discussion in #800 (comment) @paulfantom I guess we need to decide whether we want these types of alerts in mixin and on what severity level :) |
|
Right now, this will not work due to the cadvisor bug. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Helps with #759, second attempt of #760. Also may be related to #112 and may supersede #800?
In Kubernetes 1.24, kubelet started exposing a metric that counts OOMKill events for specific containers, container_oom_events_total, which I used for this alert.
This alert will fire if there are any of these OOMKill events in a container. Multi-process containers like webservers that have multiple "worker" process could silently be OOMKilled without this. I have personally seen a pod running Gunicorn throw a 100% error rate due to OOMKills that 1) didn't show up in app-level monitoring, since the workers died before recording stats, and 2) didn't show up in any existing
kubernetes-mixinalerts since PID1 never died.IMO this alert might be better than #800 since it's more granular (at the container and process level). The
OOMKilledpod status may be incorrect since it just checks ifexit_code == 137, which is caused by any SIGKILL, not just the OOMKiller.Open to suggestions!