Werk #19134: OTel: Correctly discover already running otel-collector process
| Component | Site management | ||||
| Title | OTel: Correctly discover already running otel-collector process | ||||
| Date | Jan 22, 2026 | ||||
| Level | Trivial Change | ||||
| Class | Bug Fix | ||||
| Compatibility | Compatible - no manual interaction needed | ||||
| Checkmk versions & editions |
|
The otel-collector init script did not correctly detect an already running otel-collector process when running omd start or omd start otel-collector multiple times. As a result, it attempted to start a new process and overwrote the file tracking the currect process id (pid).
The newly started otel-collector process then failed due to the ports already being used by the running otel-collector instance. This caused the original process to become orphaned, and omd status incorrectly showed the otel-collector as stopped.
To work around this issue, users had to manually kill the running otel-collector process before restarting it with omd start otel-collector.
This Werk improves the init script to correctly detect the running otel-collector process, preventing the start of duplicate processes and avoiding overwriting the process id tracking file.
After updating, users can safely run omd start or omd start otel-collector multiple times without causing failures or orphaned processes, enhancing reliability and ease of use.
If the otel-collector is active and used in checkmk users should check if their service is running correct with omd status. If the service is running no further action is required.
If the service is stopped but data is still being received, switch to the site user via sudo omd su <SITE> and run ps ux | grep otelcol.
Note down the process id and run kill -9 <process id>. Run ps ux | grep otelcol and make sure the process is stopped.
Restart the process with omd start otel-collector and wait till it is started correctly.