We can see the "catalis-assessment" service is currently in a FAILED state and is connected to the repo ravsteve/CATALIS-UPLOAD-12-14-25 on the main branch. If the "Connected branch does not exist" message persists after disconnect/reconnect, try removing the GitHub App installation entirely at github.com/settings/installations, then re-install it and re-connect the repo and branch in your service settings. You can also try creating a fresh service and connecting the repo there to see if the issue is specific to this service's configuration.

Hi there, could you please once more try removing the entire GitHub App installation entirely at github.com/settings/installations, then re-install it and re-connect the repo and branch in your service settings?

The GitHub App reinstall partially worked — other services in the project (catalis-website staging) are now triggering from pushes. But catalis-assessment/production is still not triggering. The branch shows main with no error in settings. Something specific to this service's webhook registration is broken.

Please provide a direct link to the service in question.

catalisleadership.com

assessment.catalisleadership.com

Hello!

I took a look at the branches appeared for me. Are you able to try logging out -> in to railway and see if those branches still fail to appear?

The issue is not branches not appearing. The issue affects all production services. Git pushes to main do not trigger auto-deploy for any production service (catalis-assessment or catalis-website). The only thing triggering is a failed attempt on catalis-assessment in staging, which doesn't even have an assessment service. Production deployments are completely broken for all services. Please investigate why production webhooks are not firing.

This was the third Railway employee on this thread. Each one asks one question and moves on without resolving anything. I am a Pro customer and this has been broken for weeks. I need someone to actually investigate and fix the webhook registration for production deployments — not ask me to try things I've already tried. Please escalate to someone who can look at the backend and get this working. This is a railway problem, not a problem in my end.

Hey Steve,

The way it works is that we have a rotating support on-call system. What complicates matters further is that GitHub has upstream errors when it comes to linking their service. So when it happens when you report it, it hard for us to re-pro.

The only known work around is unlinking and relinking. We do have the product engineering team working on this issue but it's a bit of a needle in a haystack situation, sorry about that.

[Broadcast] Hey, Angelo here. I'm using a tool to reach out to everyone reporting this issue at once. We've been investigating why auto-deploys and PR environments silently stop triggering, and have traced it to upstream GitHub API requests not being handled gracefully on our end. We've shipped additional monitoring to measure the false positive rate and pinpoint exactly where the issue lies. We'll follow up once we land a core fix. In the meantime, disconnecting and reconnecting your repo at the service level (Settings > Source Repo) is the best workaround. Sorry for the trouble.

Angelo, thank you for reaching out to me. It is helpful to know that I am not the only one, and that you are working on it. Just so you know, I have tried disconnecting and reconnecting endless times, and it still does not work. The only workaround I have successfully done is to use Railway Up CLI. I have been really happy with Railway so far, but if this isn't resolved, it may force me to use a different service. I do not want to do that. I have loved using Railway. I really hope you can figure this out. I would be very grateful if you would keep me in the communication loop going forward as you deal with this. Thank you.

I merged something that will get me more logs to look into it.

Quick debug here, when you make a new service does it error?

Creating a new service connected fine, no error. The issue is only with auto-deploys on my existing service.

Thanks for testing that. and on the new service, new events go through yea?

Seems like its what I think it is pertaining to webhook registration.

After I added and deleted that new service as a test, CLI deploys are now also broken. railway up uploads successfully (compressed, uploaded, build log URL generated) but no build is triggered. The UI redeploy workaround only redeploys old code. GitHub webhook pushes also not triggering. Both deploy paths are broken. Need this fixed urgently.

I see the recent CLI deploy attempts are failing because the service has a root directory configured as "assessment-app" but that directory is not found in the uploaded code. This is a separate issue from the auto-deploy problem. You can check and update the root directory setting under your service's Settings > Source > Root Directory.

The root directory is set to assessment-app — that's correct and hasn't changed. CLI deploys work fine. The issue is only with GitHub auto-deploy and deploying from the Railway dashboard.

Angelo — I appreciate you staying engaged on this and I understand it's a difficult bug to isolate. Is there any way to please have this escalated or prioritized so someone can dig into the logs directly and provide a resolution timeline? I'm relying on this for a production app and have been working around it long enough.

We've looked into this on the backend, and the Railway GitHub App installation does not currently have the correct permissions for your repository. This is something that would need to be corrected on your end in your GitHub App installation settings.

This doesn't add up. Angelo on your team previously confirmed this is a known bug on Railway's end affecting multiple users — upstream GitHub API requests not being handled gracefully. I made no changes to my GitHub App installation or repository permissions. The deploy worked continuously from day one until it stopped, with no action on my side.

Can you reconcile this with Angelo's prior diagnosis? If the answer has changed, I need to understand what changed and why. If it is a permissions issue, I need specific instructions for exactly what setting to change and why it would have broken spontaneously without any action on my part.

I'm on a deadline — conference Wednesday April 9. CLI deploys are working as a workaround but I need auto-deploy resolved or a clear escalation path.

The Railway GitHub App installation does not currently have the correct permissions for your repository. This is something that would need to be corrected on your end in your GitHub App installation settings.

Brody yiou keep repeating yourself but i checked GitHub and the permissions are in place and again, this porbkem just started suddenly. i need help here or i will be forced to move to a dfferent site o handle my deployments. i cannot cintinue klike this and i have been offered zero help so far. why am i paying for support?

Hey Steve, I'm sorry for the runaround. I'm looking at your service backend-side right now to check the webhook registration state directly. Will follow up shortly with what I find.

Thanks. I will wait for your update.

We've re-checked on our end and confirmed there are no platform-side issues affecting your account. The Railway GitHub App installation for your repository does not currently have the correct permissions. GitHub App permissions can change without direct user action, for example during GitHub App updates or permission scope changes. You would need to verify and re-authorize the permissions in your GitHub App installation settings at github.com/settings/installations.

it does have proper permissions. I just confirmed that for the nth time. Your refusal to offer any real substantive support to a paying customer is frustrating. I am moving my site to Digital Ocean by this weekend at latest if you do not help me resolve this. YOur lack of support is inexcusable, and says you do not care about your customers. I will also be posting reviews that say the same.

We understand the frustration, and we're sorry for the experience. The fact remains that in every case we've investigated, including yours, the root cause has been incorrect GitHub App permissions. This is something that exists outside of our platform and is not something we're able to fix on our end. We'd recommend reviewing the Railway App's permissions at github.com/settings/installations to ensure the correct repository access and permissions are granted.

I have reviewed that repeatedly, and all permissions are correct.

We owe you an apology, both for the contradictory responses and for how long this has dragged on. To be clear, this is a known platform-side issue affecting many users, not a permissions problem on your end. Our engineering team is actively working on a fix with an open escalation. We'll follow up on this thread when we have a resolution. In the meantime, CLI deploys via railway up remain the most reliable workaround. Sorry for the trouble.

I appreciate your acknowledging that it is not a permissions issue. It was quiote frustrating being told that over and over when i can see om github clealry that the permissions and have always been correct. Yes, I hve been using railway up and that is working for now. I am anxious to resolve this, but willing to manage with that workaround for now, as long as I know you are working on a solution and not just shoving me off with comments about permissions. Do you have any guess as to a timeline for resolving this?

We don't have a specific ETA yet, but this is actively being worked on by our engineering team. To clarify, it does appear to be a permissions issue. There are some users who have been able to resolve it, but we don't yet understand why. We will let you know when we have a solution.