-
Notifications
You must be signed in to change notification settings - Fork 24
Open
Description
We have an issue that our bot becomes unresponsive after a restart of mattermost's reverse proxy.
Steps to reproduce:
- setup up mattermost behind nginx reverse proxy; connect bot through the reverse proxy. Both mattermost and nginx are running inside docker, with http/s ports directly published from the container.
- start bot and connect to mattermost instance (verify it's working).
- restart the nginx server on mattermost (
docker-compose restart nginx
)
What we'd expect
- The bot tries to reconnect to mattermost
What happens
- we see a log entry showing that the connection was closed
- however, the bot doesn't attempt to reconnect
- the bot also doesn't exit, process stays active
- there are no ping/pong message log entries anymore in this state
- a restart of the bot is required to solve the issue
From the log:
2023-03-16 20:03:06,274 DEBUG websockets.client % sending keepalive ping
2023-03-16 20:03:06,274 DEBUG websockets.client > PING 37 8d 3d 0e [binary, 4 bytes]
2023-03-16 20:03:06,277 DEBUG websockets.client < PONG 37 8d 3d 0e [binary, 4 bytes]
2023-03-16 20:03:06,278 DEBUG websockets.client % received keepalive pong
2023-03-16 20:03:17,949 DEBUG websockets.client ! failing connection with code 1006
2023-03-16 20:03:17,955 DEBUG websockets.client = connection is CLOSED
2023-03-16 20:03:17,956 DEBUG websockets.client x closing TCP connection
(no related messages in the log afterwards)
Metadata
Metadata
Assignees
Labels
No labels