tempo and tempo-worker momentarily go into error state during "start" hook

[michaeldmitry]

Bug Description

Its observed on multiple occasions in the CI, that sometimes tempo or tempo-worker go into error state during the "start" hook.

To Reproduce

Not much on how to reproduce it, but its been observed happening in test_workload_traces in prometheus and loki's integration test.

Environment

juju 3.6
microk8s 1.28

Relevant log output

N/A

Additional context

No response