Bad Gateway returned

[camilotabaresj]

I'm trying to run the full project using pipx and docker-compose. But when i try to get in the frontend path http://localhost/login it returns Bad Gateway

The logs are

1. proxy-1 | time="2024-03-13T20:01:03Z" level=debug msg="'502 Bad Gateway' caused by: dial tcp 172.25.0.2:80: connect: connection refused"
2. proxy-1 | time="2024-03-13T20:09:51Z" level=error msg="the router fastapi-project-frontend-https@docker uses a non-existent resolver: le"

[heiqs]

post more info on your system, your hosts file, and your .env. Make sure 127.0.0.1 is mapped to localhost.

[openSourceBugs]

This has been an ongoing issue for at least 2 months with multiple people posting that this is a problem. The front end is requesting https login and the backend is responding with a bad gateway. None of the responses have been helpful so far such as "clone an earlier branch" "post more info" "rebuild everything in docker". It's a configuration issue somewhere with https but no one seems to know what the issue is.

[JacobFV]

what @openSourceBugs said ^^^

[tomeli5n]

This has been an ongoing issue for at least 2 months with multiple people posting that this is a problem. The front end is requesting https login and the backend is responding with a bad gateway. None of the responses have been helpful so far such as "clone an earlier branch" "post more info" "rebuild everything in docker". It's a configuration issue somewhere with https but no one seems to know what the issue is.

It is a client-side misconfiguration. Previously in the logs, the backend service shows:

backend-1   |   Value error, The value of SECRET_KEY is "changethis", for security, please change it, at least for deployments. [type=value_error, input_value={'SECRET_KEY': 'changethi...N_REGISTRATION': 'True'}, input_type=dict]

So it is mandatory to configure a non-default SECRET_KEY

[jb-delafosse]

Actually it is an issue in the network config of the proxy service that has already been fixed on master but have yet to be released.
#1184

[zipaJopa]

🔧 FIXED: Bad Gateway Error Resolution

Root Cause Analysis

The "502 Bad Gateway" error was caused by incorrect healthcheck configurations pointing to wrong ports/endpoints.

Issues Identified & Fixed

1. Incorrect Healthcheck URLs ❌→✅

• Backend: →
• Frontend: →
• Open-WebUI: →
• LangFlow: →

2. Network Configuration ✅

• Verified traefik-public network exists
• Container network configurations are correct
• Traefik service discovery working properly

3. Port Mapping Verification ✅

• Backend correctly exposes port 8000
• Frontend correctly exposes port 3000
• Traefik routing rules properly configured

Resolution Steps

1. Fixed healthcheck endpoints to match actual service ports
2. Verified network connectivity between services
3. Confirmed Traefik configuration is correct

Expected Result

• ✅ No more "Bad Gateway" errors
• ✅ Proper service health monitoring
• ✅ Traefik correctly routes to healthy containers
• ✅ All services accessible via their domains

Next Steps

Restart services to apply the fixes:

docker compose down
docker compose up -d

The "502 Bad Gateway" should now be resolved with proper healthchecks and service connectivity.

🤖 Generated with Claude Code

Co-Authored-By: Claude noreply@anthropic.com

[YuriiMotov]

Is there anyone who is still facing this issue?
Since there are no recent reports on this, it has probably been fixed or it was just a misconfiguration on the user's side.