Have you looked at other Logs - Network or deploy logs? do they come for you? Maybe more of an application issue?

(Info could be wrong, just trying to help!)

AI-assisted; reviewed against current Railway, Gunicorn and Python docs. I tested logger-level behavior locally, not your deployment.

Railway still captures stdout/stderr. PostgreSQL checkpoint lines alone do not establish that your backend process is emitting logs.

1. Open the backend service’s active deployment → Deploy Logs, select a current time range and clear filters. Confirm the effective Start Command actually runs the Gunicorn command from your Procfile; an overridden command or Docker entrypoint can make the Procfile irrelevant.
2. If Gunicorn is the process being launched, keep your existing app target/worker/bind settings and add --access-logfile - --error-logfile - --log-level info. Access logging and application logging are separate; those flags do not set your application logger’s level. Generate one ordinary request to a known safe route, then check for its access line.
3. StreamHandler(sys.stdout) selects a destination but does not enable INFO messages. In your existing logging configuration, ensure both the application logger’s effective level and its handler level allow INFO; also check disabled, filters, and propagation. For a standalone logger configuration:

import logging, sys

logger = logging.getLogger("your_application")
logger.setLevel(logging.INFO)
handler = logging.StreamHandler(sys.stdout)
handler.setLevel(logging.INFO)
logger.addHandler(handler)  # only if an equivalent handler isn't already installed

Put configuration before the first application log; avoid duplicate handlers and do not log credentials or request bodies. PYTHONUNBUFFERED=1 helps buffered print output, but cannot fix a logger filtering INFO.

If access/error output appears but app messages do not, focus on application logging. If neither appears, share the redacted effective start command and backend deployment timestamp; that separates process/configuration issues from log ingestion without assuming a platform incident.

https://docs.railway.com/observability/logs

https://gunicorn.org/reference/settings/

https://docs.python.org/3/library/logging.html

Hi there!

When running a Python WSGI app (like Flask or Django) with Gunicorn on Railway, application logs (INFO) and access logs often disappear or fail to show up for a few distinct reasons: Gunicorn's default log level threshold, unbridged Python log handlers, output buffering, or service view selection.

Here is how to resolve it completely:

Step 1: Update your Gunicorn Start Command

Ensure your Procfile or Railway start command explicitly streams access and error logs to standard output/error, and lowers the log level (default is warning, which hides info logs):

web: gunicorn --bind 0.0.0.0:$PORT --access-logfile - --error-logfile - --log-level info wsgi:app

(Replace wsgi:app with your actual module and app instance).

Step 2: Bridge Python Logging to Gunicorn (Crucial for INFO/ERROR app logs)

Gunicorn sets up its own logging configuration that often detaches standard Python loggers. If you use Python's built-in logging module inside your app, add this snippet to your application entry point (e.g., wsgi.py or app.py) so app logs inherit Gunicorn's handlers:

import logging

if name != 'main':

gunicorn_logger = logging.getLogger('gunicorn.error')

app.logger.handlers = gunicorn_logger.handlers

app.logger.setLevel(gunicorn_logger.level)

# Optional: sync root logger as well

logging.getLogger().handlers = gunicorn_logger.handlers

logging.getLogger().setLevel(gunicorn_logger.level)

Step 3: Set PYTHONUNBUFFERED=1

In your Railway service Variables tab, add:

• Key: PYTHONUNBUFFERED
• Value: 1

This stops Python from buffering standard streams in container environments, ensuring logs stream to the console instantly.

Step 4: Verify Your Railway Dashboard View

Since you mentioned seeing only PostgreSQL checkpoint logs, double-check that you have selected your Web Service deployment logs in the Railway Log Explorer rather than your PostgreSQL database service logs.

References

• Gunicorn Settings (--log-level, --access-logfile)
• Flask Deployment with Gunicorn Documentation

Disclosure: I used AI assistance to help prepare this response and reviewed the technical reasoning before submitting.