Per-instance metrics for multi-instance services | Feature Requests

Per-instance metrics for multi-instance services

planned

Anurag Goel

For multi-instance services, it can be helpful to see per-instance CPU/memory metrics to identify bottlenecks.

July 28, 2020

Andrew Duca

This would be an awesome feature

Gerry Agbobada

Would that mean exposing the internal IPs as well for each instance? If I want to use a Prometheus instance to scrape service metrics on top of accessing node-level metrics, just accessing the replicas through the load-balancer in

hostport

is not enough for my use case

Anurag Goel

marked this post as

planned

Aleff Souza

Anurag Goel I think this could also be applied to the logs, having all logs from multi-instances in a single "terminal" is quite confusing.