Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

APM high CPU usage?

I have an 8.5.1 slx system that was monitoring ~5500 elements. This server was fine (slx, custom, AM, netflow (~10), syslog and trap) until I uninstalled AM and added APM. Before (and I am back there now) this servers CPU ran ~30%. No major gaps, nor any slowness issues (remote sql and website). But after I installed APM, CPU usage shot up to ~90% with the poller looping between running and stopping.

I called Tech support, and they informed me that I needed a second poller, but I now do not believe so. Remembering I had no trouble until I installed APM, I removed APM and am now back to ~30% CPU load. The funny thing is, I was averageing ~90% with *No* Applications setup yet. Just sitting there, the JobMonitor.exe service was taking ~ 50-60 % of the CPU with nothing defined. I did not convert any of my old application data (as a matter of fact, I truncated both tables).

Has anyone else seen this? I would like to use APM, but my current machine cannot handle it at idle, let alone with any applications defined. I feel that this application should not use ~50% by itself without any apps being monitored. To me this would have to be a bug. My budget this year does not include an extra poller, but what would an unconfigured APM be pulling 50% itself?

Thanks,

MarkS.

Find more posts tagged with

Accepted answers

All comments

mark_wiggans

APM SP3 addressed a High CPU issue= Do you have this applied?

josh.clark

Which service/process is using all the CPU? We don't have one named JobMonitor.exe.

salyerma

Ok, I found out which one it was. SolarWinds.JobEngineService.exe.

But here is the interesting part. I had two of them running when I had APM installed. I remember that distinctly. Curious part though was that only one of them was taking up CPU utilization. It would take up the bulk of the CPU.

So I went and looked again, they are still installed. JobEngineService and ScheduleService are still installed as a service after I removed APM. There were a couple on uninstalls/reinstalls with APM, but I just kinda figured that was normal to have more than one instance running.

Yes, SP3 was installed.

Marks.

EDIT: Ok, I just reinstalled APM and I have to correct what I was saying. I only have one JobEngineService.exe in the task list, one SchedulerService in the task list but I have *3* JobEngineWorker.exe's listed. I looked through the forum and found where you guys said to delete the extra services that get left over. I seem to be more stable now with less cpu usage. I will watch it tonight.

No sooner had I posted this, then 2 more JobEngineWorker's show up. I have a total of 3 with each one taking about 20% util. This is without any app monitors configured.

josh.clark

It is not unusual to have multiple JobEngineWorker processes. The Job Engine service creates these processes to do the actual polling for a monitor. You should ever see more then 3 or 4. It doesn't make sense that the worker processes would show up if there is no work to do (i.e you have no application configured).

I wonder if your database go into an odd state for some reason. If you open SQL Studio or Orion Database Manager and execute each of these queries, do any rows get returned?

SELECT * FROM NPM_ScheduledJob
SELECT * FROM NPM_ActiveJob

If so, it would seem things got out of sync somehow. You could reset things by doing this:

1. Stop the SolarWinds Orion Job Engine and SolarWinds Orion Job Scheduler services

2. Execute this SQL

truncate table NPM_ScheduledJob
truncate table NPM_ActiveJob

3. Start the SolarWinds Orion Job Engine and SolarWinds Orion Job Scheduler services

Does that help any?

salyerma

Ok, that was it. I ran both of the select commands, and there were ~3500 records in the scheduledjob and ~1500 in the activejob tables. Now I am back down to ~30% CPU with APM installed.

Couple of things. When I tried to remove AM to install APM, AM could not be removed ( I had to rebuild from scratch keeping the DB). The conversion process (AM - APM job conversion) would not run for me (constant missing job ID). Even after removing APM and reinstalling several times, I wonder if the jobs were leftovers from when the conversion would fail? I *never* had that many monitors in AM. The website for APM never showed any jobs configured and no way of knowing anything was going on. Maybe happening to others if it is happening to me.

Second thing. Just some constructive criticism, but Phone Support instantly tried to sell me a second poller. Actually have a sales person calling me to talk about a demo of the second poller. Now this is not to say that we would have found out a second poller would not have helped, and he was giving me a demo to make sure, but within two emails you guys had it solved. Maybe my situation can help support help others

But I will try adding a couple of monitoring templates and see what happens.

Thanks,

MarkS.