cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Agent Unavailability and Performance Warehouse Data

Babar_Qayyum
Leader

Dear All,

How we can plot the measurements data on the chart in case the agent/host is not available?

The use case is, that in the last week load testing was performed on two of the DR applications and those applications were temporarily connected to our UAT environments so once activity was done then the DR application machines were shutdown.

As per the default configuration of AppMon the process/host will be removed from the Agents Overview dashboard after 72 hours so due to this the process/host will not be available for the splitting/plotting purposes.

Now the management/business is asking to make the comparison reports between UAT and DR but we do not have process/host to map the PWH data until we turn on the DR machines once again.

Looking forward for your valuable advises to sort out the situation.

Regards,

Babar

8 REPLIES 8

Joe_Hoffman
Dynatrace Champion
Dynatrace Champion

Babar,

If I understand your use case, the time series metrics should still be in the PW, even if the agent no longer shows up in Agent Overview.

If you put the measure on a chart, then select a timerange for the chart that includes the period when the agents were active, then there should be values on the chart.

Does this not work?

Hello @Joseph M. H.

Thank you for your answer. I was doing a basic mistake that from the beginning was trying to find the process in the measure splitting.

What can be the consequences in case we increase the time from 72 hours to more for these type of exercises?

Regards,

Babar

The only issue you might encounter by going beyond 72 hrs is that you go back in time before you started collecting this measure.

Also, keep in mind that data is digested down so we don't keep the same resolution of data for older data. Details of these values is in the doc. But this is generally not a problem as if you're looking at data over a 6 mos time range, you don't need high resolution data. But if you go back 6 mos ago and expect high resolution data, it won't be there because we don't store high resolution data that far back. Again, the doc shows the details.

https://community.dynatrace.com/community/display/DOCDT65/Performance+Warehouse+Administration?_ga=2.173827875.140498702.1521474972-485961544.1518021892

Hello @Joseph M. H.

Apologize I could not understand the first point 'The only issue you might encounter by going beyond 72 hrs is that you go back in time before you started collecting this measure'.

Can you please more elaborate, because I am really looking for the consequence before doing this activity in the UAT environment?

The PWH resolution is very well understandable and we have already configured the required resolution settings.

Regards,

Babar

Sorry for the confusion.

The "Monitoring" dashboard is a special dashboard designed to 'simplify' the view of operational data. However it's based on the same source of metrics that are stored in the PW. So the data is in the PW, but the Monitoring dashboard itself does not provide you a view into that data before 72 hrs, but the data is still in the PW. To view it, simply build your own dashboard and chart whatever metric you want.

Please ignore my comment about going back in time before you started collecting the metric. My point was simply to show that if you started collecting a metric on 1-Feb, then there simply would not be any data to chart before 1-feb.

HTH

joe

Hello @Joseph M. H.

I know that the conversation is getting prolong a bit 🙂

Basically I am looking for the increasing the index time of the agent which is by default 72 hours in case of unavailability of the process/host.

What can be the consequences in case of increasing the time?

Regards,

Babar

Babar,

Can you first explain how you planned to change the index time of the agent to something other than 72 hrs?

The 72 hr limit is based on the default cache optimized in-memory storage of metrics and is not directly controllable by the customer. But even this isn't relavent as any query that exceeds the in-memory data range simply causes a PW query. So the whole optimization is transparent to the user.

Hello @Joseph M. H.

To be very honest, I am not aware about the time increase and I was thinking that will ask you about the flag later on.

Thank you for the explanation of the 72 hours limit which is very helpful for my knowledge.

I was reading about the Unique agent names with index feature from the below link and was confused that maybe we can increase/decrease the in the same manner.

https://www.dynatrace.com/support/doc/appmon/installation/set-up-agents/unique-agent-names/

Regards,

Babar