[plugin-barman-cloud] Readiness probe does not allow HA

[faymard]

Hi,

We are using the plugin-barman-cloud chart to use object storage for backups. For HA requirements, we increased the replicaCount to 2 or 3, and we notice that inactive pods do not open port 9090. They open it when they achieve lease acquisition, but until then they are marked as down, which triggers false positive alerts and noise.

Is there anything we can do to mitigate this ?

Regards,

Code :

charts/charts/plugin-barman-cloud/templates/deployment.yaml

Line 76 in da9471f

tcpSocket:

[dosubot]

Hi @faymard! I'm Dosu and I’m helping the charts team.

This is a known limitation with the plugin-barman-cloud chart: only the pod that acquires the leader lease opens port 9090, so inactive pods remain unready and trigger false positive alerts in HA setups. This behavior is hardcoded in the deployment and operator logic—there’s no chart setting or documented workaround to change it right now. The readiness probe is a TCP check on port 9090, which only passes after lease acquisition, so only one pod is ever "ready" at a time, and the rest are marked as down until they become leader deployment.yaml.

Mitigations are limited: you can adjust your alerting rules to ignore non-leader pods, or run only a single replica to avoid the noise. There’s ongoing discussion about this in the community, but no fix or configuration option is available yet issue 419 issue 687.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[tuunit]

Will be fixed with #776