Open OchiengEd opened 1 month ago
/assign
@OchiengEd, what do you think about instrumenting the Operator Controller with the OpenTelemetry SDK? So, we could expose metrics, traces, and logs in a standard way.
The goal of the metrics in this case is to detect drastic changes in compute resource utilization during development. Therefore it should be able to be run from CI.
Attached is the baseline metrics report on the current state of compute resource utilization for operator-controller
and catalogd
.
Scripts and Jupyter Notebook used to generate the above report can be found in the olmv1-stats repository.
Gather baseline metrics for catalogd and operator-controller to be used as a reference point.