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
The thread memory feature does not work well when using multiple replicas.
Implementing leader election and forward inbound requests to the leader could solve this issue.
The text was updated successfully, but these errors were encountered:
I think it is easier to add leaderelection, which is a kubernetes standard than add a redis client in your app, and make every query read or write to the redis. But I'm not a skilled dev. You can find example of that, in kyverno source code, for instance.
From a "user" perpective, I would prefer that calert use election than having to deploy any additional redis pods.
From an architecture perpective, one would want to run redis in ha, to avoid a spof. This is getting too heavy for calert.
If a spof is acceptable, then we can run calert on a single pod.
The thread memory feature does not work well when using multiple replicas.
Implementing leader election and forward inbound requests to the leader could solve this issue.
The text was updated successfully, but these errors were encountered: