Current etcd monitoring documentation only mentions the /metrics endpoint (prometheus) to collect metrics. Metricbeat module uses the v2 stats endpoint.

They seem to offer complementary metrics (there are some metrics that I couldn't find in /metrics endpoint) but the main problem I see is that v2 API doesn't seem to cover v3 store and gRPC operations, and the /metrics endpoint seems to cover both stores and gRPC.

etcd 3.3 includes an experimental v2 emulation layer over v3 stores, we could try to collect v3 stats using this feature, but this would require users to enable non-default experimental features, what is not ideal.