retry content signature requests on error from autograph

[jcristau]

Sometimes when trying to get a content signature from autograph we get a 502 (or other error?), we should retry them a couple of times to try and reduce the errors that propagate to clients.

[gbrownmozilla]

The request is made by sign_hash(), at https://github.com/mozilla-releng/balrog/blob/main/src/auslib/util/autograph.py#L15

[g-k]

we get a 502 (or other error?)

Yep, when an autograph instance is down or unhealthy ELB will return a 502.

I looks like RelEng signingscript uses a retry_async helper to retry for all exceptions https://github.com/mozilla-releng/scriptworker-scripts/blob/5a1ef390a2727f9d7c28bc089fb2a890c90b2c3f/signingscript/src/signingscript/sign.py#L960