Schedule Database Maintenance for Additional Polling Engine (APE)

Background Issue:

Interface graph statistics very slow to show and mostly timeout this experience include NTA as well.

slow or mostly timeout showing graph

Troubleshooting:

We have 2 APEs and this timeout issue mostly occurs on objects monitored under APE-2

From MPE always showing this warning from DB maintenance log, this also shows when you run the active diagnostics.
Run DB maintenance on MPE and running DB maintenance plan on SQL server not helping, log is still showing same timeout experience

2025-08-09 02:33:02,132 [1] WARN  SolarWinds.Orion.Common.SqlHelper - Long Running Query: ALTER INDEX [PK_APM_HardwareItem_Hourly] ON [APM_HardwareItem_Hourly] REORGANIZE Execution time 77 seconds

I found the issue is mostly under APE-2 I tried to check if the server has DB maintenance program like in MPE, and IT DOES!!!
Checked the log of "swdebugMaintenanceExecutionHistory.log" it never runs so this confirms there is no task scheduler or trigger running this program.

Tried to run it, I saw log index defragmentation and bla bla.. (I also run on APE-1 as well)

Directory: C:\ProgramData\SolarWinds\Logs\Orion\swdebugMaintenanceExecutionHistory.log

After waiting both finished and test again the interface graph and NTA.

NOW IT"S WORKING FAST !!

Issues:

- I don't know why DB maintenance plan on SQL server it's not affecting even though it has reorganize index and rebuild index. Is there any schedule recommendation to run this maintenance plan? daily or weekly?

- How to create schedule Database maintenance program on APE? I'm looking to schedule the same time with MPE which is daily on 2:15AM

Find more posts tagged with

database maintenance

Database

npm

Additional Polling Engine

slow

ape

nta

slow performance

Accepted answers

Lofstrand

The SolarWinds Database maintenance task does several things. Like aggregating data, cleaning up old things. So it must be ran, but only from one polling engine, PE.

Index defragmentation is also done by the SW DB Mnt Task, but that's optional. (And if you ran that from several places it doesn't hurt but will probably don't do any good.)

On your settings page, under "Polling Settings", you can adjust if maintenance should include index defrag and what time to run the whole job.

If you implement index defrag yourself, yes you may do it like you suggest - or the way I suggested. If you are not a dba or don't have one inhouse, maybe your way is a bit easier. A dba would choose another way. All three ways should work, it's just a matter on how much control you want.

Suggest you go on with the "Maintenance Plan Wizard" that you started with, it can't hurt :-)

All comments

Lofstrand

As there is only one database for all polling engines, only one of the should need to be running database maintenance and that should be the MPE.

If you try to run the job from several PE's at the same time, the would probably "collide" and not create a faster/better job.

Why did it run better after DBMaintenance from an APE? Don't know. But could be that index defragmentation job was better performed at another time. You could try to add your own index defragmentation jobs on the SQL server, see if that helps. You will also have better control over what indexes are defragmented. If you don't have any preference on how to maintain your SQL databases I would suggest looking at SQL Server Backup, Integrity Check, Index and Statistics Maintenance Ola Hallengrens solution.

moonb0w

so DB maintenance is executing same thing? if that so that is unnecessary then

I created DB maintenance plan using wizard on SQL, does it different with the link above?
sorry I'm not a DBA
I create the plan refer to this guide from Solarwinds added with backup DB (Full)

How to Create a SQL Server Maintenance Plan

Lofstrand

The SolarWinds Database maintenance task does several things. Like aggregating data, cleaning up old things. So it must be ran, but only from one polling engine, PE.

Index defragmentation is also done by the SW DB Mnt Task, but that's optional. (And if you ran that from several places it doesn't hurt but will probably don't do any good.)

On your settings page, under "Polling Settings", you can adjust if maintenance should include index defrag and what time to run the whole job.

Suggest you go on with the "Maintenance Plan Wizard" that you started with, it can't hurt :-)

moonb0w

thanks @Lofstrand appreciate for your help

yer

Just as curiosity, what version do you have?

moonb0w

i'm on 2024.2.1