Skip to content

Add healthchecks for the services #26

@sunu

Description

@sunu

Problem description

While running the vector service through the helm chart, it failed to startup with the following error:

 .... <snip></snip> ....
  File "/usr/local/lib/python3.11/asyncio/tasks.py", line 479, in wait_for
    return fut.result()
           ^^^^^^^^^^^^
asyncpg.exceptions.InvalidPasswordError: password authentication failed for user "s00pers3cr3t"

ERROR:    Application startup failed. Exiting.

This error should ideally kill the parent uvicorn process but it doesn't. It was still running:

root@vector-695c469fbd-qfxlg:/tmp# ps aux
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root           1  0.0  0.3  31056 26228 ?        Ss   14:52   0:00 /usr/local/bin/python /usr/local/bin/uvicorn tipg.main:

We don't have any explicit liveness probes either. So to kubernetes the pod appears healthy:

$ kubectl get pods -n eoapi | grep vector
vector-695c469fbd-qfxlg                     1/1     Running   0          22m

Expected Output

The pod's liveness probe/ startup probe should fail and kubernetes should know that the pod failed to start or has died.

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions