semantic PR check still not working

[dev-nicelee]

The semantic PR check doesn't work since yesterday.

The status shows that it is waited for the check. Even after the branch up-to-date, the issue is still going on.

There was no update in semantic.yml config as well.

[jamesmbourne]

I'm seeing the same for PRs created recently (within the past 30 mins).

[pamelasarkisyan]

I also experience the same since Monday.

[zeke]

Hey folks. Sorry to hear it. I haven't made any changes so I'm not sure what's up here. I will try to investigate today.

In the meantime if you're blocked, I'd suggest standing up your own instance of this app, or using a GitHub Action to do your semantic PR checks. See the README of this repo for more details about that.

[andypattenden]

Looks to be working now

[pamelasarkisyan]

Update: We were using semantic-pull-requests v2.3.0 and GitHub Enterprise 3.1.14 which was fine, but after the upgrade to GitHub Enterprise 3.3.1 and app stopped working as shown in the screenshot of this issue. After deploying a version from master branch the app worked fine.

As the last official release of this app was over 2 years ago it might be a good time for another release 😊

[zeke]

Thanks for that bit of info @pamelasarkisyan.. I didn't realize the release process was broken. semantic-release is a bit finicky.

I'm not sure releases are really related to this issue though, as the hosted version of this app is deployed to Heroku whenever the default git branch is updated. 🤔

[zeke]

Looks to be working now

Good to hear. Working for others too? I didn't change anything.

[dev-nicelee]

Thank you, it works now😊

[Haroenv]

I'm actually seeing it broken on multiple repositories, even on recently created PRs. I guess we'll need to look into hosting our own instance

EDIT: I think it works again now

[gilmatok]

@dev-nicelee Please reopen this issue as this is sporadic.

Going over the code, it doesn't seem like an issue with the code itself, but rather an issue with actually running this bot.

However, there doesn't seem to be any way to debug GitHub checks...

[chrislambe]

Also still seeing this issue, though somewhat intermittently. Very strange!

[t-kelly]

I just installed the app, and I'm not seeing a check show up on my PRs. Is this related?

[jkantner]

@t-kelly Yeah. The check status remains yellow when stuck. Renaming the PR multiple times though will trigger the check. We shouldn’t have to do that.

[gnavalesi-navent]

We're also seeing this issue but not in all repositories. All have the same configuration:

commitsOnly: true
allowMergeCommits: true
allowRevertCommits: true
anyCommit: true

In some repos the check status remains yellow, even if we change the title or add more commits

[dev-nicelee]

Probably this is not the best solution for this issue, but I found that the issue has been resolved by deleting semantic.yml file. Maybe this is the clue for this, @zeke?🤔

[ianchudson-caribou]

We saw this issue for ~30-45 minutes today where it was failing. It seems to have just come back online without any changes on our end.

Is there a viable way to debug these types of issues where it appears to be an external issue?

[zeke]

Hey folks. Thanks for being polite and trying to help discover the source of the issue. I took a look at the Heroku app that's running this service today, and it's just been falling over a lot, presumably from increased usage, as no changes have been made to the app itself recently.

I went ahead and bumped it from a Hobby Dyno to a Standard 2X, which has more memory, CPU share, and "compute".

Let's see if that helps. 🤞🏼

For posterity, here's a snapshot of recent app metrics:

[gilmatok]

@zeke Thanks, Zeke. I'll keep you posted on this issue if I see any improvements.

[zeke]

Looking better so far!

[gilmatok]

Haven't had any issues so far! @zeke

[dev-nicelee]

@zeke Thank you so much! I think I can close this issue unless the issue emerges.🤔

[zeke]

Thanks @dev-nicelee. Closing!

[guilhermetod]

Experiencing this hanging now

[zeke]

Thanks for reporting @guilhermetod -- this should be fixed now.

It was a Node.js upgrade that passed CI but failed in production: https://github.com/zeke/semantic-pull-requests/pull/177