Why Is the Runner/Forwarder Replica Count Set to 1 in Helm?

[jcastill-akamai]

Hello team,

I'm curious why the Helm replica count is hardcoded to 1. Is there a specific reason for this? I'm asking because we want to ensure high availability and would like to overwrite these values ideally without modifying the official Helm chart.

Would increasing the replica count affect how the runner or forwarder components operate? For example, if we scale the replicas, will multiple instances of the runner/forwarder triggers when a playbook its met? Our scenario involves nodes that may be terminated, so we want to make sure Robusta continues running on at least one other node if that happens.

Forwarder: https://github.com/robusta-dev/robusta/blob/master/helm/robusta/templates/forwarder.yaml#L12
Runner: https://github.com/robusta-dev/robusta/blob/master/helm/robusta/templates/runner.yaml#L10

I appreciate any insights you can share.

[jcastill-akamai]

i see this its cover already on #419 although i see this its from 2022 so wondering if its still on the roadmap

[aantn]

Hi @jcastill-akamai, we're not currently focusing on this! We're more focused on HolmesGPT, which does agentic investigations (vs deterministic playbooks) and was built to scale better.

Can you share more on your use case? Maybe there is a way HolmesGPT could help there instead?