Reputation: 9542
I'm running a flask application on Heroku and a few times a week I see errors like the following:
pskb-prod heroku/router: sock=backend at=error code=H18 desc="Server Request Interrupted" method=GET path="/administrator/index.php" host=tutorials.pluralsight.com request_id=d1331164-89f2-4e01-b0d3-0666fa4541ad fwd="85.102.23.202" dyno=web.2 connect=0ms service=409ms status=503 bytes=18571
I'm assuming this is some kind of hacking attempt to see if that URL exists and works. Obviously it doesn't since I'm running a flask app, not PHP.
Why does this show up as an H18/503 error?
I can try the request in the browser myself to /administrator/index.php
and correctly get a 404 response from the flask app.
How does this H18/503 error get triggered?
It's worth noting this has nothing to do with the specific URL because I see similar errors for related URLs like:
/admin/login.php
/bitrix/admin/index.php?lang=en
/admin
/wp-login.php
Finally, I don't see any of my dynos being restarted around the time I get these errors either so don't think it's related to the dyno rebooting at just the time these requests come in.
Upvotes: 3
Views: 482
Reputation: 9542
I think the issue is related to gunicorn without nginx as https://github.com/benoitc/gunicorn/issues/840.
Upvotes: 1