DX Application Performance Management

Hi,

We use APM 9.5.2 and dashboards/metrics in investigator are not showing any data in Webview over a period of time.

Our webview is running on same server on MOM server and it has 4 GB memory allocated. I understand we did not follow proper recommendations from CA. But I wanted to know exactly what is causing webview to behave like this. From logs, I get only below minimal information to arrive the root cause. Our MOM is running fine and displaying dashboards. So no issue in MOM and cluster.

I am thinking to enable webview UI monitor to get some more information to analyze the response time but we are waiting for approvals to enable.

[WebView] Timeout while retrieving data for dashboard

BTW, can some one tell me urgently what could cause this type of errors? I find difficulties in analyzing webview performance and most of the articles I see are related to core APM architecture. Where do find ready made performance tips for webview?

Thanks,

Karthik

Any inputs?

Hi,

as a first step I would recommend setting up monitoring for WebView: https://wiki.ca.com/display/APMDEVOPS98/How+to+Configure+and+Monitor+APM+WebView+Performance#HowtoConfigureandMonitorAPMWebViewPerformance-ConfigureAPMUIPerformanceMonitor

Any other errors in the WV.log or EM.log? How "big" are your dashboards? Live or historical, number of metrics?

Ciao,

Guenter

Hi Karthik,

Are you able to compare the WebView performance in 9.5.2 against earlier versions of WebView, that you notice a drop in performance ?

The presentation layer changed in 9.5 to use HTML5 compliant technology as opposed to relying on Silverlight in the past.

One thing we had seen is that dashboards can be slow to load where there are a lot of alerts on the screen.

This is because the default behaviour is to 'draw' all of the alerts in the background and then 'paint' them to the screen when they are all ready.

So that can be a reason for the timeout.

It would be interesting to know if you have issues on dashboards where there are little or no alerts posted.

We've not necessarily seen the same problems in the Investigator so it's not necessarily true that your cluster is perfectly fine, there could still be something that could be tweaked on that side, particularly as Guenter says if you might have large amount of historical data, or perhaps metric groupings and alerts in dashboards are evaluating a large amount of metrics.

Thanks,

David

Guenter_Grossberger - Yes, we are in process of enabling webview UI monitor as first step. There is no other errror in webview log or EM log. The dashboards are just having four graphs trying to pull data. So it's not too big.

DavidLewis I did not do a comparison of performance with earlier versions. To be specific, it's not only dashboards. We started seeing the issue in dashboards not having any data. Later on we saw there are no data displayed anywhere in webview. No data available for metrics in investigator too. Only agent name, host name, metric category, metric names are seen. When we click on particular metric, it tries to pull data but ended up with blank graph.

I am not seeing any useful messages in logs as well.