Long-running query SQL process fragment Unknown

fifthrace

Since upgrading to 3.1, I've noticed that on the Long-running query alert my SQL process fragment is no longer showing as it did before in the UI.

I apologize if I missed something that stated why this would be. Any help would be much appreciated.

Thanks all,

Robbie

Chris Spencer

Apologies but this is a regression in SQL Monitor v3.1.

A bug has been raised (ref: SRP-6952) and we'll be looking to fix this for the next release.

Regards
Chris

EdCarden

What if we downgrade back to the prior version of 3.0.X we were using, can we get back the misssing SQL Fragment info?

I hope this bug has been given a high level of priorty because the LRQ (long Running Query) alert is just about useless without this key piece of information.

I'm suprised that this is not being fixed till the next release of SQL Monitor. Then again maybe no other user besides myself makes use of the LRQ alert and so this is not a major setback/bug for anyone else.

Thanks

Chris Spencer

This issue is fixed internally but we're still in the process of scheduling our next release. I'll speak to some people and if possible get back to you with an estimated release date.

Downgrading to 3.0 would get the SQL Fragment info back but I'm not sure if the v3.0 software would work with a v3.1 data repository. I'd have to test that.

Regards
Chris

EdCarden

Chris,

FYI - So far the downgrade is working. I have yet to have any probelms or errors and most importanty, I got back my missing SQL Fragment info for LRQ alerts.

I was worried that the info had been permanently lost and not just missing but its still there. I just wish I knew how to get that info directly from the DB.

Thanks

Chris Spencer

That's good to know that a downgrade appears to work.

Regarding getting the information out of the repository, I've had a little go at doing this and come up with:

USE RedGateMonitor -- or whatever the repository name is
GO
SELECT utils.TicksToDateTime(us.CollectionDate) AS CollectionTime           
      ,us._Command
      ,us._CumulativeProcessorTime
      ,us._DatabaseName      
      ,us._FullCommand      
      ,us._Status      
  FROM [data].[Cluster_SqlServer_SqlProcess_UnstableSamples] us
  JOIN [data].[Cluster_SqlServer_SqlProcess_Keys] spk
  ON us.Id = spk.Id
  WHERE spk.[_SessionId] = 99 -- replace this value with Process ID from the alert details page.  
  ORDER BY CollectionTime DESC

I think that this at least maps SQL fragments to Process IDs.

Regards
Chris

Saint007

I got the same problem. Let me know when this issue is fixed and release.

JKG

Can we get an estimated date for the fix availability. I use this frequently and need to decide about rolling back.

Chris Spencer

We're planning to release SQL Monitor 3.2 next week, most probably on the Tuesday or Wednesday (August 7-8). This will contain the LRQ fix.

I think that this is a fairly safe estimate but if it changes I'll post an update.

Regards
Chris

Chris Spencer

Just to let you all know, v3.2 has been released and can be downloaded from:
http://www.red-gate.com/sqlmonitor

This release fixes the LRQ alert issue.

Regards
Chris

fifthrace

Thanks Chris, seems to be working better now.