Action should pretend a "cache miss" on networking issues

[foxylion]

Description:
Underlying infrastructure issue for GitHub artifacts cause a lot of issues for workflows depending directly or indirectly on actions that use the caching infrastructure from actions/toolkit. See actions/cache#820 for details.

I would highly suggest that this action should be more resilient regarding service outages in the backing caching infrastructure.

The issues of today would most probably not be visible if the cache action would do "nothing" (like a cache miss) when the caching backend can't be reached or the server responds with an incorrect answer.

The @actions/cache action does show how a resilient implementation can look like: https://github.com/actions/cache/blob/main/src/restore.ts#L55

Currently there is now exception catching done in the @actions/setup-node action: https://github.com/actions/setup-node/blob/main/src/cache-restore.ts#L44

Justification:
It is unacceptable that problems with GitHub caching/artifact infrastructure will effectively make most workflows unusable.

[IvanZosimov]

Hi, @foxylion 👋 ! Thank you for the feature request! We will definitely take a look at that problem.

[IvanZosimov]

Hi, @foxylion 👋! This PR mitigates the problem. By now I'm going to close this issue. If you have any additions feel free to ping us 💌