Zammad-web.service Randomly Stops Responding - No Errors in Logs

Technical assistance
1

Hi Everyone,

I’m encountering an issue where the zammad-web.service process on my server randomly stops responding. Restarting the service using systemctl restart zammad-web.service resolves the problem temporarily, but the issue reoccurs unpredictably.

Details:

  • OS: Debian 12
  • Zammad Version: 6.4.0-1732871355.611ba7e2.bookworm
  • Behavior Observed:
    • The service suddenly becomes unresponsive.
    • Checking journalctl -u zammad-web.service shows no errors or warnings at the time of failure.
    • Nginx logs indicate a series of HTTP 499 errors right before the issue occurs.

Example Nginx log entry:

XXX.XXX.XXX.XXX - - [06/Dec/2024:13:19:34 +0100] “GET /api/v1/users/search?term=a.s HTTP/2.0” 499 0 “https://xxxxxxxxxx.eu/” “Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36 Edg/131.0.0.0”


### Troubleshooting Steps Taken:

* Restarted related services:
  * `nginx.service`
  * `elasticsearch.service`
  * `postgresql.service`
None of these resolved the issue.
* Only restarting `zammad-web.service` fixes the problem temporarily.
* Checked logs for Zammad, systemd, and Nginx but found no actionable errors.
* Monitoring server resources (CPU, RAM, etc.) does not reveal any apparent bottlenecks.

Does anyone have insights into what might be causing this issue or suggestions on how to debug further? Could the HTTP 499 errors from Nginx be related?

Thank you in advance for your help!
TaKeN
2

See Daily timeout (error 504) - #2 by MrGeneration

Please search the forum next time.

3

Thank you for the response.

I had reviewed the mentioned thread before posting and while my issue seems similar, I believe it’s not the same problem. In my case, there are no timeout errors in the logs:

root@zammad:/var/log/nginx# grep “upstream timed out” zammad.access.log
root@zammad:/var/log/nginx#

However, I do see HTTP 499 errors in the Nginx logs.
Additionally, in my case, the issue is resolved by restarting only the zammad-web.service, without needing to restart whole zammad.

If you have any further insights or suggestions, I’d greatly appreciate your guidance.

4

They are the same image.