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

This product reached the end of support date on March 31, 2021.

Limitation of Calculate Results

Babar_Qayyum
DynaMight Leader
DynaMight Leader

Hello Dears,

Is there any limit to add the number of measures in the business transaction?

Regards,

Babar

13 REPLIES 13

Hello Babar,

Just tried to add an hugh amount of measures to a business transaction, and surprisingly i got an 'save profile' error, may be there is a limitation although we don't recieive a warning.

----------------------------------------------

An amount of 100 worked for me, this must be enough right? 🙂

Do you also get an error if you add a hugh amount of measures? If yes may be you can create a support ticket for it.

----------------------------------------------

From dynatrace server log:

[StoreSystemprofileRequestProtocol] Failed to deserialize system profile: com.dynatrace.diagnostics.server.communication.requestprotocol.systemprofiles.StoreSystemprofileRequestProtocol readRequest:45

Regards

Hello Steven,

Thank you for your prompt reply.

The situation is that we deliver hundreds of MQ/MW services (Native and JMS).

We filtered the services with the destination queue:/// and adding the measurements of count and time for each service to plot the charts for the MQ/MW teams.

At present on a certain level, it isn't allowing me to add further more because I already added more than 100 🙂

What can I do to accommodate all the services in the single business transaction instead of creating business transaction for each service?

Your expert opinion can make my life comfortable.

Regards,

Babar

Hello Steven,

🙂 I was able to add 126 and then I go the same message.

Regards,

Babar

Hello Babar,

Try to understand your situation, you have a message bus which is consumed by different teams? Each team has his own consumer service which you want to monitor?

---------

I think the best practice is to have a business transaction split measure who represents all your services, and then add count and response time to the calculate measures.

----------

Question is how to obtain the split measure? (if someone has other possibilities please let us know)

You depend on your code structure, if your message bus consumers have a common execute task method with the MQ/MW service name as an argument, you could place a sensor & "method arguments value measure" on that argument. If you then use that measure in a BT as split measure you will have a dashboard like this:

Service one (teamA) | Count10 100ms

Service two (teamB) | Count20 200ms

Service three (teamC) | Count30 300ms

(in a way of speaking you moved all your calculate measures to one BT split measure)

-------------

Is it first possible to check below performance clinics? It will give you a big insight in measures / sensor and business transactions.

Business transactions

https://www.youtube.com/watch?v=HdCqPuFCfOQ

Measures measures measures

https://www.youtube.com/watch?v=h1ornaNgzpU

Advanced business transactions

https://www.youtube.com/watch?v=BLz7qc5tstU

Hope this helps, if you have more questions let us know.

Hello Steven,

Thank you for your suggestions and helping to resolve this situation.

I attached a below screenshots for more understanding about MW Services and our requirements:

Le me explain more about the calculate results, if you see there is a service called Account Transfer Service which is using by an application and this service has a request and reply.

Our requirement is to observe the response time with the load so we included the count as a load on the application.

The plotted chart look like following:

Now like this service we have to configure 100 of services and each service has minimum 2 or 3 measures of time and count.

Once I started work to fulfill this demand, on that time I never realized that on one stage calculate results will be hindrance.

Now I came too far 🙂 and feeling helpless.

Regards,

Babar

Hallo Babar,

Can you add a picture of the measure details (right click edit), the details of the filter measure and all the calculate measures.

Thx

Hello Steven,

Below are the screenshots of the filter measures and the calculate results measures:

Please let me know if any further information required to handle this situation and I would like to ask you that meanwhile we are working to achieve this target should I open the ticket with DynaTrace to get their thoughts about this?

Regards,

Babar

Hello Babar,

Got it! Thx for the extra clarification, we definitely should have a measure who contains all the destination names and use that measure as a split result into the business transaction.

Found below url into the documentation

https://community.dynatrace.com/community/display/...

A first try: Could we use this measure as a split result?

Destination/MsgType – The destination and message type of the sent or received message. Each row shows messages grouped by destination and message type.

Regards

Hello
Steven,

I
had gone through with the messaging dashboard and found that there are two
following common names after the each MW service destination name but the
services have distinct names:

/WebSphereMQ
Message

/JMS TextMessage

The
below screenshot and the highlighted ( in yellow) as an example that what
information we are looking for measurements and charting purpose e.g.

Acct_Bal_Req/WebSphereMQ Message


  • Count (Load)
  • Time (Request Time)

Acct_Bal_Rply/WebSphereMQ Message


  • Time (Reply Time)

Acct_Bal_Rply/JMS TextMessage


  • Time (Overall Response
    Time)

I guess this should clear the actual concept of desired business transaction.

Regards,

Babar

Hello Babar,

I am glad you have a solution for your issue,

Some extra info

I think that the messaging dashboard works op top of the session storage (contains purepath data), that's perfect for rather small time periods. If you want bigger timeperiods you can count on the performance warehouse. This is a database (sql server, oracle, post gresql, ...) with aggregated data. How to do it? You can create a business transaction with the Destination and messagetype as split result together with count, response time, ... as calculate measures. Be sure that the 'store results to the performance warehouse' setting is active. A business transaction hotspot dashboard or a graph (with your BT as input) will show you the same message data, but this time you can also select big time periods. Obviously session storage dashboards have other advantages compared to performance warehouse dashboards, but this is out of the scope of this forum question.

Regards

Hello Steven,

We already have calculate results for the count and response time for each service and you know the purpose of these measures to plot the chart and create the incident for each service.

Can you give me an example in the visual mode e.g. screenshots of the proposed BT? So I can understand exactly your point and also what I am missing.

Regards,

Babar

Hello Babar,

Below performance clinic will give you a lot of info on how to make incidents

https://www.youtube.com/watch?v=Ysh_HL8HDoA

As for the graph is something like this, each line could be a duration / count for one destination message name. You also will see the destination message name into a legend.

Regards

Hello Steven,

Thank you for your continuous assistance.

I would like to open an informational ticket with Dynatrace Support to understand the limitations of BT Calculate Results.

Regards,

Babar