Slow Response on WebServer
AdamMetcalfe
Posts: 11
Hi support,
I am getting slow response when browsing to the site for SQL Monitor. Sometimes it does not even respond.
IIS Server + Base Monitor on an IIS7 Windows Web Server 2008 R2
Database is on a SQL2005
Client is IE 8
Number of servers being monitored is 4
I notice that the server is taking a lot of memory on web server.
Thanks!
I am getting slow response when browsing to the site for SQL Monitor. Sometimes it does not even respond.
IIS Server + Base Monitor on an IIS7 Windows Web Server 2008 R2
Database is on a SQL2005
Client is IE 8
Number of servers being monitored is 4
I notice that the server is taking a lot of memory on web server.
Thanks!
Comments
RedGate.Response.Engine.Alerting.Base.Service
We are currently investigating an issue where the base monitor machine will suddenly start using large amounts of memory. We believe it's being caused by a combination of trace being enabled for the monitored server and the number of alerts being generated. As each alert is generated we query the server for the relevant trace data for that point in time. We are currently looking to see if there is any inefficiency in the way we do this.
Some more details about your problem could be very useful to us
1. Whilst the memory issue was happening, how many alerts were you getting for the monitored server that had trace enabled?
2. Does the memory issue occur when trace is enabled but the most commonly occurring alerts have been disabled or had their thresholds reduced so as not to be triggered as often?
3. Was the monitored server under stress at the time and likely to be generating huge amounts of trace?
In the meantime I can only suggest that SQL Monitor's trace option is used purely as a diagnostic measure and not turned on permanently.
Many Thanks
Chris
Test Engineer
Red Gate
The install is essentially straight out of the box and as such configuration of the alerts etc has not changed, from the defaults.
The servers were not under any extra load when the trace was run. As such no extra alerts were being raised when trace was on.
I note today that the memory being used by the service is now at 462mb.
And CPU is al reduced on the server.
I hope this helps, if i can provide extra information (logs etc) please let me know.
Well that's good news that the base monitor memory usage has stabilised at a more acceptable level but we appreciate that it seems to be down to trace now being disabled.
Issues of this nature are notoriously difficult to diagnose but it's possible the SQL Monitor log files could contain information. On Server 2008 they would be located at "C:\ProgramData\Red Gate\Logs\SQL Monitor 2". Only log files generated whilst the memory usage was high would be likely to contain useful relevant information however.
If you experience this issue again then we'd be interested to know the number of handles RedGate.Response.Engine.Alerting.Base.Service is using. This can be viewed in Task Manager (the column isn't visible by default but can be added using View > Select Columns > check Handles). If time allows then the sysinternals tool "handle.exe" (very small download from http://technet.microsoft.com/en-us/sysi ... 96655.aspx) logs very useful information about performance. It's a command line tool that can be ran as follows:-
handle.exe -p RedGate.Response.Engine.Alerting.Base.Service > handleinfo.txt
handle.exe -p RedGate.Response.Engine.Alerting.Base.Service -a > detailedhandleinfo.txt
Again this would only be useful to us if the service is performing badly at the time.
Any files (log file or handle.exe output) can be zipped up and sent to me at chris.spencer@red-gate.com.
Many Thanks
Chris
Test Engineer
Red Gate