The proclogger tool logs the creation, removal and renaming of processes.
For example, it enables you to:
- Check if your cronjobs are started the way you expect them.
- Verify that your maintenance daemons are running the way you expect them (e.g. PostgreSQL autovacuum).
- Educate your audience to never provide passwords as command line arguments. (If you demonstrate this issue with
top, your audience might think "oh, it appears just shortly, perhaps nobody will notice." But a logging tool makes the problem obvious.)
This tool is for you, if you:
- Often run
topjust to observe which processes are started and how long they run.
- Decrease the refresh time of
topdown to the sub-second range, so you don't miss short running processes.
- Try hard to read the
toplist fast enough before items vanish.
- Try hard to observe processes that change their command line at runtime.
At the moment, this tool polls over
/proc with maximum speed. While this approach is very simple and straight forward, it has two serious drawbacks: high CPU usage and yet the risk to miss parts of fast changes.
Any pull requests to implement a more efficient and more reliable approach are welcome.