Add a metric that tracks max image scaler processes
We currently track the number of running image scaler processes, but not the threshold set by max_scaler_procs
. By also exporting this piece of configuration to Prometheus, we will be able to compute a component saturation metric to see how close we are to exhausting actual / max
.
It should probably also supersede concurrency_limit_exceeds_total
, or at least be favored over the latter for monitoring and alerting, but I haven't quite decided on this yet. Let's track both for now.
Related links:
Edited by Matthias Käppler