I'm having this problem too https://station.railway.com/questions/postgres-redeploy-appears-stuck-pg-boun-8530ee6c

Hey everyone, posting the full solution since I had multiple issues

stacking on top of each other. Hopefully this helps someone else.

Symptoms:

• Build failing on deploy
• Gunicorn WORKER TIMEOUT every 5 minutes
• "Unable to connect to database via SSH" in the dashboard

Root cause #1 - Python attestation (build failure)

mise 2026.6.11 introduced attestation verification for GitHub artifacts.

If your runtime.txt is pinned to an older Python version (e.g. 3.10.13),

the build will fail.

Fix: update runtime.txt to python-3.13.5

Root cause #2 - flask db upgrade hanging at startup

The Procfile runs flask db upgrade before gunicorn. If Postgres isn't

ready yet, it blocks indefinitely and the build times out.

Fix: replace the direct command with a retry script (migrate.sh):

#!/usr/bin/env bash

set -e

MAX_ATTEMPTS=5

DELAY=2

attempt=1

until flask db upgrade; do

if [ "$attempt" -ge "$MAX_ATTEMPTS" ]; then

echo "[migrate] Failed after $MAX_ATTEMPTS attempts."

exit 1

fi

echo "[migrate] Attempt $attempt failed. Retrying in ${DELAY}s..."

sleep "$DELAY"

DELAY=$(( DELAY * 2 ))

attempt=$(( attempt + 1 ))

done

Then in your Procfile:

web: cd backend && bash migrate.sh && gunicorn app:app --bind 0.0.0.0:$PORT

Root cause #3 - gunicorn workers timing out

Even after the build, if Postgres is slow to respond, the app's seed/init

code blocks the worker past gunicorn's timeout (default 30s).

Fix: add connect_timeout to your SQLAlchemy engine options:

if database_url:

engine_options['connect_args'] = {'connect_timeout': 5}

This makes connection failures fast (5s) instead of hanging forever.

All three fixes together got the app back online. Good luck!