I'm doing fullstack development on my home machine but I also have a (private) github repository which is connected to Heroku. I am generally happy with Heroku as a host except the title.
The site runs fine, then after about 10minutes of testing out the embedded quiver editor, it freaks out and the database goes away. Here is the tail of my Heroku log:
2025-07-22T06:11:42.677961+00:00 app[web.1]: self._connect(self._config.default_access_mode)
2025-07-22T06:11:42.677961+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.12/site-packages/neo4j/_sync/work/session.py", line 130, in _connect
2025-07-22T06:11:42.677961+00:00 app[web.1]: super()._connect(
2025-07-22T06:11:42.677962+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.12/site-packages/neo4j/_sync/work/workspace.py", line 161, in _connect
2025-07-22T06:11:42.677962+00:00 app[web.1]: self._pool.update_routing_table(
2025-07-22T06:11:42.677962+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.12/site-packages/neo4j/_sync/io/_pool.py", line 802, in update_routing_table
2025-07-22T06:11:42.677962+00:00 app[web.1]: raise ServiceUnavailable("Unable to retrieve routing information")
2025-07-22T06:11:42.677963+00:00 app[web.1]: neo4j.exceptions.ServiceUnavailable: Unable to retrieve routing information
2025-07-22T06:11:42.678480+00:00 app[web.1]: 10.1.88.200 - - [22/Jul/2025:06:11:42 +0000] "GET /db/edit-content/af1175584d7d43e7bbcc2b44f5733144 HTTP/1.1" 500 123663 "Sign In" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/138.0.0.0 Safari/537.36 Edg/138.0.0.0"
2025-07-22T06:11:42.679546+00:00 heroku[router]: at=info method=GET path="/db/edit-content/af1175584d7d43e7bbcc2b44f5733144" host=arrowglue-dd35800b80e8.herokuapp.com request_id=a34f2b84-303d-9359-4443-f0d3d9855cf7 fwd="98.97.140.210" dyno=web.1 connect=0ms service=121ms status=500 bytes=123663 protocol=http1.1 tls=true tls_version=unknown
This also seems to happen on the local side as well, I've been getting "Unable to retrieve the routing information" for a few days.
Anyone had this error? I don't have it when working with Neo4j desktop (locally).
It is a cryptic error, but looks like a session expiring on the heroku side (when you click "Sign In" it takes you to heroku, and the stack trace also has herokuapp.com).
Doesn't seem to have anything to do with Neo4J.
1 Like
Actually, I'm getting it both locally and on live Heroku site. So the common denominator their is the AuraDB neo4j database connection.
So... I'm not sure what to do, I don't seem to be getting any errors from this...
Would be glad to add an expert as collaborator to view and run the code on GitHub. Since we're using AuraDB, all you would need is your Python IDE, Python installed and to run the pip -r on requirements.txt for example.
Normally, routing is a client side issue ... ?
e.g. "I can't reach my destination" ... the problem isn't the destination - it is either the client's configuration or the mechanism to reach the destination.
If you can connect to your aura instance from your local machine's browser ... then it is a configuration problem locally.
1 Like
Yes, but everything is working fine until about 10 minutes of up time... Then BOOM
... then is a session timeout/logout.
1 Like
@joshcornejo Yes, I'm also getting SessionExpired afterward. But that causes the whole app to break i.e. I think restarting the dyno on Heroku fixes it. That's not right... a session timeout of Neo4j is not even 10 minutes - its an hour. So, not sure what is timing out 
2025-07-22T06:11:42.677962+00:00 app[web.1]: self._pool.update_routing_table(
2025-07-22T06:11:42.677962+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.12/site-packages/neo4j/_sync/io/_pool.py", line 802, in update_routing_table
2025-07-22T06:11:42.677962+00:00 app[web.1]: raise ServiceUnavailable("Unable to retrieve routing information")
2025-07-22T06:11:42.677963+00:00 app[web.1]: neo4j.exceptions.ServiceUnavailable:
@joshcornejo In my above log paste is the above. This looks like a "connection pool manager" is doing something and fails, like it's an internal library issue.
If you are getting SessionExpired then you probably sat idle for too long ... and you only get the error after being logged out/session expiration.
It is technically not an "application error" ... you have been logged out and you can't execute whatever you are trying.
1 Like
If I get this in code, and it's capturable. How can I restart the "Django neo4j service"?
I think you are asking a heroku question in a neo4j forum ...
1 Like
The Django session doesn't timeout, meaning I'm still logged in with Allauth (Github account), on the site. It's the Neo4j connection that's timing out...
Heroku has no forums. I can ask: AuraDB (which I'm paying for) but that's here! Django, or Neomodel forums. I've asked in Django just now.
But this still seems highly relevant place to post.
