Job somtimes starts too late

Hello,

I have a job that has to run every 5 minutes (It's a short runner and take only 1-2 seconds to complete). In most cases the job runs intime but sometimes it's started a little bit too late.
"Too late" means over 1 minute delay in rare case more that a 2 minute delay. I found no other jobs running at the same time when this occures.
Do you have an explanation why this could happen?

Master is a Linux server with 4.1 multiplatform. I can see the timestamp in scheduler.log on the master.
I also cannot see any heavy load on the master that would prevent the job from running in time. I do not expect to run the job exactly with a "00"-sceond-timestamp. But a 1 or 2 minute delay seems a little bit too long.

Any hints?

Thanks in advance

Does the timestamp match actual job start time?
How big is the schedule.log file?

>> Does the timestamp match actual job start time?
Yes, it does.

>> How big is the schedule.log file?
scheduler.log ist 600k
debug.log is 100k
velocity.log is 10MB

Strange that you ask about the logfile size. How could this effect whether a job is started with a delay or not?

I'd suggest trying to clean schedule.log or simply rename it. I suspect that a large log file with long job messages can make the server busy when updated HTML logs. In you previous posts it sounds like you are using HTML reports option. That processing might be causing delays on every job start or stop. After you rename the log file, either use Log Clear menu in 24x7 or restart 24x7 if it is running as a "service", just to ensure it doesn't rebuild the log file. Please let us know if renaming the log file makes any difference.