Counter metrics in GCP metrics explorer

Question

I have a DataFlow job with a counter metric. On every restart the metric is reset to zero, as expected. The problem is that when using the counter in gcp Metrics explorer, I cannot get an accumulated value for the metric, disregarding restarts. Prometheus has a function called increase() that does this. Is there a similar function for gcp metrics explorer?

Answer 1

One approuch to metrics across runs would be to make use of Cloud Monitoring. There is a good how to on the features and usage of custom metrics.

If you use job names that you can apply a regexp to then you can make use of the filters to aggregate them into a graph.