This message was deleted.
# rke2
a
This message was deleted.
b
I got sharded Prometheus partially deployed - 5 shards now...
but it means I only get 1/5th of my metrics 🙂
g
We replaced the rancher-monitoring application with the kube-prometheus-stack as it had many bugs. As for this many pods, you need to have a long list of metric drops, but that still could overwhelm Prometheus. We're currently looking into VictoriaMetrics to overcome this. It also has a "stack" helm chart to install.
for metric drops look into the tsdb-status page of Prometheus and start weeding out the metrics you don't need.