Trusted publisher workflow: server produces 500 with action version 1.8.6

[deathbots]

In testing out the new GHA trusted publisher workflow, the

Trusted publishing exchange failure: 
Token request failed: the index produced an unexpected
500 response.

Example failure

The job that fails with the 500

Strong possibility the error is on my side, but the 500 indicates otherwise of course :) Please let me know if you see something obvious or if I can try something new.

[webknjaz]

@woodruffw ^

[woodruffw]

Thanks for the ping @webknjaz!

@deathbots: is this a regression with the latest release of the action, or is this the first version you're trying? Also, could you share a screencap of your trusted publisher setup?

(I can't triage the 500 directly, but @di or another PyPI admin can.)

[di]

Hey @deathbots, I tried to leave you a note here: runZeroInc/runzero-sdk-py#11

This should be resolved now!

[deathbots]

Thank you @di you are our hero! Whatever you did, it worked great. And thank you very much to all in thread for getting the OpenID Connect work over the line and into hands.

[webknjaz]

@di FYI I've just got a 503 too.

@woodruffw it's weird that the annotation only shows this single line:

Trusted publishing exchange failure:

In the logs, though, I see:

Notice: Attempting to perform trusted publishing exchange to retrieve a temporary short-lived API token for authentication against https://test.pypi.org/legacy/ due to __token__ username with no supplied password field
Error: Trusted publishing exchange failure: 
Token request failed: the index produced an unexpected
503 response.

This strongly suggests a server configuration or downtime issue; wait
a few minutes and try again.

(https://github.com/ansible-community/ansible-pygments/actions/runs/4951889218/jobs/8857581936#step:4:28)

I wonder if it's time to implement some sort of an exponential backoff...

[di]

Hmm, is the 503 actually coming from the token exchange request, or is it from the upload request? Our object storage provider is currently having a minor outage which I think will manifest as a 500-level failure here, but should only affect the upload, not the token exchange.

[woodruffw]

Based on the error message, it looks like it happened during the token exchange (that's the only time we render the HTTP status code like that).

@webknjaz I'll have some time to look at fixing that annotation render tomorrow 🙂

[woodruffw]

I'll have some time to look at fixing that annotation render tomorrow

I think this is a linebreak thing: the message in question is wrapped over multiple lines, so only the first gets the annotation.

[woodruffw]

Opened #156 with the annotation fix.