Few questions / statements on logging. I'm troubleshooting slowness since installing SEUM on my standalone APM box. The majority of this slowness seems to be on the APM side but I'm still trying to determine if SEUM has a part of it.
1. Logging from remote agents seems to be lacking on the main SEUM install. I had one playback on a remote agent fail and it wasn't clear why. SEUM reported the playback as having not played back yet. To me it seems like there needs to be some better information on the main console about why a job might be failing without having to resort to logging in to the remote agent. I previously used the remote logs succesfully to fix an issue so the data is all there it just seems like it needs a phone home option to get some better detail into the console.
2. I'm getting a bunch of warnings on long running queries. Possibly related to perf issues I'm having with APM as well.
2011-08-29 16:02:35,603  WARN SolarWinds.Orion.Common.SqlHelper - Long Running Query:
SELECT t.TransactionId,t.RecordingId,t.AgentId,t.Name,t.Description,t.Frequency,t.IsLicensed,t.JobId,t.LastStatus,t.LastDuration,t.LastDateTimeUtc,t.WarningThreshold,t.CriticalThreshold,e.ServerName AS PollingEngineName
FROM SEUM_Transactions t
INNER JOIN SEUM_Agents a ON t.AgentId = a.AgentId
INNER JOIN Engines e ON a.PollingEngineId = e.EngineId
WHERE t.TransactionId in (5)
Do I need to be concerned with these?
to answer your questions:
1. Remote agents log only locally. We could transfer logs from agent to main somehow but in majority of cases it would be just unnecessary network bandwidth usage. To troubleshoot remote agent you need to go to that machine. On main you can see if job was correctly sent to agent but that is all.
2. Long running query may be caused by APM that is using database a lot. That particular query you mentioned is nothing complicated so there is really no reason why it should take a long time by itself.
Have you tried to unmanage all SEUM transactions to see if performance issues are caused by SEUM?
i ended up moving SEUM to a dedicated server. This alleviated a great deal of the issues I was having in APM.
However it created new challenges such as displaying my SEUM information on my APM maps...
can you provide more details about your APM and SEUM issues? What did you find out regarding this? What were the issues you had with APM? What were the issues you had with SEUM?
If SEUM can cause some issues with APM we definitely want to know about this so we can have a look.
Some interesting notes on the issues I've been having.
As I mentioned before to resolve my APM issues I moved SEUM to a dedicated server running off the shared APM db server with its own db. No issues in either product.
That was going so well for a few days that I thought I'd retry the SEUM / APM combo. APM got slow immediately once I loaded SEUM to the APM server even w/o any transactions loaded ( aside from the previous data that existed in the db).
SEUM in this configuration is sharing the APM DB.
Not sure what this means but I will attempt to leave it in this config for a little while in case I can provide you guys any meaningful information.
I'd prefer to keep APM and SEUM on the same machine for ease of mapping but this may not be possible.
thank you for update.
Can you find out what part of infrastructure is slow when APM and SEUM are together? Check if IIS process eats CPU, if it's database that can't keep up with the load etc? If there is not single point of slowness, it will need more investigation, probably via a GoTo meeting.
There are some diags on case#267870. The earlier one should be more accurate to what I was seeing as the second was after support had me attempt a few things.
I'm not clear if it is an APM and SEUM issue or an issue of simple oversubscription.
I do have SEUM and APM seperated now and still hitting the same DB server without issue so the contention seemed to be on the shared front end box.
SEUM was complaining of slow running queries, APM was very slow, I saw high memory in LSASS, Large APM maps failed to load at all even with increased time outs, AJAX lists were very slow or failed to load in both products.
If you guys want to take a look I can probably reinstall SEUM onto the APM box.
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process.