[helm] airbyte-bootloader can't recover from failures

### Helm Chart Version

2.2.0

### What step the error happened?

On deploy

### Relevant information

Use of a `Pod` object for the `airbyte-bootloader` component doesn't track completion, which can result in broken deployments.
In our case an incorrect Postgres endpoint was provided, but the bootloader pod was not re-scheduled on a re-apply of the chart with an updated `values.yaml`. I should note we used ArgoCD for the deployment here; `helm` wasn't directly used.

Using a `Job` rather than a `Pod` would simplify re-triggering a failed job, as the `Job` could be updated and the failed `Pod` deleted, which would result in another `Pod` with the correct settings being created.

As an added bonus, a `Job` object could better integrate with Helm hooks, with the bootloader running on the `pre-install` and `pre-upgrade` hooks. I'm not familiar with Airbyte's DB migration method, so this may not be applicable if the `airbyte-bootloader` is only intended to ever run once.

I've patched a local copy of the Helm chart, changing the `airbyte-bootloader` from a `Pod` to a `Job`. Let me know if you're open to raising a PR.

### Relevant log output

```shell

```

---
**Internal Tracking:** https://github.com/airbytehq/oncall/issues/13284