The resource state metrics provide different metrics paths depending on the cluster
Iceber opened this issue · 1 comments
Iceber commented
What would you like to be added?
The resource status metrics for all clusters are currently exposed in the /metrics
path, and although flags are provided to filter the metrics, the data for these metrics grows linearly as clusters increase.
To ensure that monitoring tools like prometheus can process metrics faster, I'd like to provide different metrics paths for different clusters, e.g. /clusters/<cluster name>/metrics
Of course we also need to make prometheus automatically discover and collect metrics from new clusters, so if you have any ideas, please comment!
Why is this needed?
It is possible to ensure that the resource state metrics of a cluster are smaller under a single path, so that monitoring tools such as prometheus can better collect them
clusterpedia-bot commented