exporter stops iterating after 500 metric types

[jotka]

I was trying to debug the way exporter works. It seems that the first REST call is the authentication, then the second one is a query about the possible metrics types, and them it iterates over the list to get the metrics

The second call gives me 3382 metric types (I'm using --monitoring.metrics-type-prefixes=kubernetes.io)

But then it stops iterating after 500 metrics types, so I don't get all the metrics I need.

Is it an expected behavior or my misconfiguration?

[kgeckhart]

Do you have anymore info on this? Looks like the default page size for the descriptors is going to be 10k so all of those descriptors should be passed to the callback. Is it possible that most of them are getting picked up by the deduping logic here, https://github.com/prometheus-community/stackdriver_exporter/blob/master/collectors/monitoring_collector.go#L261-L273?