Scaling Tempo Metrics Generator

[bmonsalvatge]

Unlike the other Tempo components, it doesn’t appear the metrics-generator has any guidelines on scaling: Size the cluster | Grafana Tempo documentation

Additionally there’s no indication if it can be horizontally scaled based on the number of replicas. And the examples in Github have it running as only 1 replica unlike the other components.

So my question would be:

1. Are their scaling guidelines for metrics-generator elsewhere in the docs that should maybe be linked back to the sizing guidelines with the other components.
2. Can the metrics-generator component be scaled horizontally.

Thanks in advance. There are two other threads on this topic that have gone unanswered (dating back over two years):

Can we run multiple instances of metrics-generator component?

Can we run multiple instances of metrics-generator component ? If yes then how the service graph processor aggregating the data? for example if instance one is getting client span and instance two is getting server span

Increasing Replicas for metrics-generator in Tempo

Hello, I have a question about the metrics-generator in Tempo. I’m running Tempo using the helm-chart. By default, the replicas for the metrics-generator is set to 1. Is the metrics-generator intended to be used as a singleton? I’m curious if it’s possible to set the replicas to 2 or more to extract metrics from a large number of spans. Thank you.

So I am hoping for some clarity on this issue.