This time we finalized and cleaned up our metrics setup within the k8s cluster, we made the Grafana dashboards compatible with the new setup (although it still seems to be a bit buggy), and we also managed to pull some metrics from Prometheus back into k8s and set up a simple auto-scaler based on it, but we still have to figure out how to set that up properly and can we even use it without it killing active player connections.
Category:
