The OpenSearch deployment gets stuck with the following error message:

opensearch                         blocked      3  opensearch                2/edge         171  no       failed to create datahub_index index - deferring index-requested event...

That blocked only happens if the index already exists, as the code shows.

After doing some investigation on data-hub's CI failure: https://github.com/canonical/datahub-k8s-operator/actions/runs/11261394134/job/31314738969?pr=5#step:13:1786

I've noticed that we may have the following sequence of events happening:

1. unit 1 is the leader, executes the index-requested event and creates the event
2. leader ship changes -> unit 2
3. unit 2 executes the index-requested event and notices the index already exists

Now, the cluster will be stuck on "failed to create {index-name} index".

It is not possible to validate the flow above as the real cause of this CI failure. Nonetheless, it is a possibility and therefore, we should avoid it.