Custom monitoring extension

[CommunityUser]

We have build custom monitoring which is schedule to run every 5 minutes. Custom tool is getting triggered every 5 minutes however Metric browser reports value per minute basis and does a dip to 0 between the execution time. We are using metric aggregation has 

OBSERVATION so it should report last submitted value. But that is not the case it dips to 0 between execution. Any idea how we can avoid the dip?

[CommunityUser]

The points that Karun listed is exactly why I am looking for interval greater than 60 secs.

[CommunityUser]

Here is the link to the documentation: https://docs.appdynamics.com/display/PRO41/AppDynamics+Essentials

Check under section 'Server Monitoring' -> 'Extensions and Custom Metrics' -> 'Build a Monitoring Extension Using Scripts'.

Thanks

[CommunityUser]

i agree, i was saying why is there a limiatation on periodic interval to be less than 60 seconds when the mtric granularity is minute.

Even if i set the periodic execution to say 30 secs, in the metric browser i will see data with one minute granulaity only, right?

[CommunityUser]

Yes, the data is rolled up to a "minute" granularity. But if you have multiple metrics that have been rolled up (lets say 6 if they ran every 10 seconds) you will see in the rollup the minimum, maximum, count, and average for all of those. So, you may find value reporting more often within a minute. Depends on your use case.

[CommunityUser]

Hi,

How are you controlling the frequency of execution? Using monitor.xml? 

I would suggest using this tag in monitor.xml 

<!--  Edit execution-style as needed. -->
           <execution-style>continuous</execution-style>

https://docs.appdynamics.com/display/PRO41/Build+a+Monitoring+Extension+Using+Scripts#BuildaMonitori...

[CommunityUser]

yeah <execution-style>periodic</execution-style>

Interval 300 secs

[CommunityUser]

is there a specific business requirement for this 5 mins interval? If you want the data to be reported every minute on the metric(atleast for 4 hours) i would suggest to go ahead and use continuous.

[CommunityUser]

Here is the deal:

Shockingly, when you use 'periodic' execution in a custom monitoring extension script, the frequency must be less than 60 seconds. This is because the agent sends data to controller every 60 seconds no matter what the aggregation type is.

From page 1508 on Pro417_AppDynamicsApplicationIntelligencePlatform_21Nov2015.pdf,

I'll try to verify this, but the above documentation appears to be clear.

Karun

www.karunsubramanian.com

APM|Monitoring|Consulting

[CommunityUser]

Karun explanation sounds good. Can you provide link to the PDF

[CommunityUser]

If the metric granularity is minute, i am not sure why would someone use a frequency of less than a minute? 

Trying to think os use case that goes along with this design.

[CommunityUser]

There are many reasons why a customer wants to run the custom monitoring script at an interval other than 60 seconds. Couple of them are:

1. He doesn't want to overload the system he is hitting with his custom monitoring script.

2. Each run of his custom monitoring script runs for more than 1 minute to complete.

Thanks