Envoy: Leak of information about cluster member

[hakuno]

Please, how do I retrieve the envoy_cluster_internal_upstream_rq_xx by cluster members?

I have a cluster with 2 members... (load_assignment/lb_endpoints)... but I can't find any information to get errors from each cluster nodes.

One of these cluster members is buggy, but I can't find which node. No tip found.

[htuch]

@hakuno does the /clusters endpoint provide you the information you need?

[hakuno]

So I can read rq_error as bad HTTP status code. Nice!

cluster::node1::rq_active::0
cluster::node1::rq_error::0
cluster::node1::rq_success::94
cluster::node1::rq_timeout::0
cluster::node1::rq_total::94

How do I read that with the Prometheus/Grafana?

[htuch]

@hakuno I think that question is outside my domain knowledge, so I'll leave this open for others to comment on for the normal question period of time (2 weeks).

[mattklein123]

Per-host stats are not currently exported to the stats sinks, due to cardinality issues.

[hakuno]

All right. Thanks!