You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Based on the comment it assumes the clustering should off because by default people don't use Loki input that needs clustering but when they do it result in duplicate logs.
We should either:
Make clustering on by default
Use the same clustering value as alloy.
Validate if alloy is configured for clustering, then clustering should also be on for alloy-logs
The text was updated successfully, but these errors were encountered:
We actually used to enabled clustering by default for all Alloy instances. Just in case that a user switches to a config that can utilize clustering. However, we turned it off in 7ec6d61 because it adds overhead for coordinating jobs that never utilize clustering.
The chart does have some guardrails to fail deployment if you change pod log gathering to something that can use clustering (gathering via the Kubernetes API), it'll tell you to enable clustering, and if you enable clustering, but are using the daemonset to gather logs from the hostpath volume mount, it'll tell you to disable clustering.
If those checks aren't working, I'd definitely like to know!
https://github.com/grafana/k8s-monitoring-helm/blob/main/charts/k8s-monitoring-v1/values.yaml#L2685-L2687
Based on the comment it assumes the clustering should off because by default people don't use Loki input that needs clustering but when they do it result in duplicate logs.
We should either:
alloy-logs
The text was updated successfully, but these errors were encountered: