feat: add Slack notifications for the post-deploy assets check

[stevejalim]

Bedrock port of mozmeao/springfield#780

[janbrasna]

LGTM.

I haven't seen the Slack notifications yet, but figured you were satisfied with how they turned out, and the sparse checkout works as expected, before porting here too. 🚢

[janbrasna]

Noticed this before, but actually prefer it this way — the main job no longer waits for the initial notifier to finish before queuing for available runner — starting the important stuff right away in parallel, no matter what slack notifier… This can easily save 10sec+ from the run time.

The only difference is before, the run visualization connected the jobs like this, depending on each other:

This one plots that disconnected, as is the reality/logic of these:

I'd deliberately keep it that way for speeding things up, just thought it's worth noting so folks are not surprised seeing the diagram for the first time.

[stevejalim]

Yeah - i wondered about chaining vs triggering them in parallel - faster is better really, and if for some reason the Slack API call fails (eg outage), we'll still run the actual checks.

Might move the others to this pattern if we find no downsides in use

[janbrasna]

+1 my thoughts exactly, so I'll simplify others to this pattern when I'm in the neighborhood…

(FWIW, Slack API action has an escape hatch that on auth, API, HTTP issues, errors, timeouts and other outages and fails it does not mark the run as failed, and only annotates, unless explicitly set to fail hard — so it should never fail the pipeline actually. But good to trigger in parallel and not depend on its results in any way…)