LOAD graph values on Tier Flow Map view

[Maria_Garcia]

Hi,

We have several tiers with several Business Transactions and several Service Endpoints (in fact almost the traffic used to be detected under Service Endpoints) but the load values are not the expected ones.

The tier flow map shows the following load:

That seems to be the Business Transactions values:

But the service endpoints of that tier are executing a lot of calls (a higher value of calls):

The Service endpoint load graph displays the expected values:

From our point of view the tier load should take into account all the calls under the BTs, I mean, also add all the service endpoints calls, but seems not be the case...

We need to confirm how dbAPM calculates the calls values displayed in the Load graph under  the Tier Flow Map view.

Which entities are taking into account to obtain that values?

Please kindly confirm it.

Thanks a lot in advance!

[Allan_Schiebold]

Hi. Since SEPs are a totally separate set of detection rules, this makes sense. The BT detection and SEP detection would have to be exactly the same for the numbers to match. 

[Maria_Garcia]

Thanks Alan,

In this case the number of calls for an specific BT (21 calls/min) is much lower than the number of executions (330 calls/min) of the SE associated to that BT.

So the load graph at tier level is not reflecting the real traffic of those nodes.

How to solve it please?

Thanks 

[Allan_Schiebold]

I would first need to see your configuration settings for your BT's as well as SEPs. Feel free to PM or post here. Thanks. If this is urgent, please reach out to the support team. 

[Maria_Garcia]

Hi,

We do not have any specific configuration, just detecting the first 2 segments and including the SOAP action

But please, what we need is to confirm and verify which values are used to calculate the LOAD graph that is displayed in the Tier Flow Map.

Thanks

[Allan_Schiebold]

Hi. So these detection rules are what determine the load and other metrics. I'm wondering if using part of the SOAP request is ignoring any transactions that don't have SOAP information. I'd have to log in to your controller to really dig in to what's going on. 

Either way, if you have an account manager they can get you in touch with your solutions engineer to look at this. Or you can open a support ticket if this is urgent. 

[Maria_Garcia]

Thanks Allan,

Which metrics are used to calculate the load graph? I mean, I do not understand why a BT with 21 calls/min have service endpoints associated with 330 calls/min in each node and the tier only reflects in the load the 21calls/min...

However other tiers in the same controller under the same rules detection are displaying values closer tan the expected ones.

Just trying to understand how dbAPM calculates the load and how to solve the issue...

Thanks 

[Allan_Schiebold]

I wouldn't be focused on the calls / min, but rather the total number of calls in the time range (which will still be different values for you in this case). 

For some reason in your controller, the detection rules for BT's is different than the config for SEPs. 

[Maria_Garcia]

Hi,

Yes we have configured it due to:

- BTs has to include the 2 segments and the soap action in order to capture the specific operation executed on the application

- SEs configuration was added because some actions are not captured as BT, it only appears as SE.

Which should be the best configuration in order to obtain a real total load in the load graph?

Thanks a lot!   

[Allan_Schiebold]

Hey there, happy Friday 🙂 In this scenario, with the current configuration and my understanding at the moment, SEPs are going to be the single source of truth. 

[Maria_Garcia]

Thanks Allan,

Yes, the SE graph seems to be the correct value but… How to solve that huge difference between BTs and SEs metrics? Which changes should we done on detection rules in order to have the right metrics on the BTs too?

We have a dashoard for support team and we see that some tiers displays the right metrics on BT load but anothers don´t (and all of them have the same configuration) so metrics cannot be used as expected...

Thanks