Question about custom metrics quota calculation

[skrystosik]

Hello,

I have question about custom metrics calculation. I have script that on start is creating custom network device, defining custom metrics and then pushing data to them. It works fine but after couple of runs I've got message that quota of 100 free custom metrics is exhausted. I've checked API and I see only 12 custom metrics. PUT of the same metrics is changing quota size every time?

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

Each Dynatrace environment (SaaS or Managed) comes with 100 free custom metrics. Additionally, 10 free custom metrics are provided for each of your allotted host units. For example, if you have 100 host units, you will be provided with 1,100 custom metrics.
100 + 10 * 100 = 1,100 (custom metrics)

There's a maximum free tier limit of 10,000 custom metrics.

Have a look in the below link for more insight.

https://www.dynatrace.com/support/help/get-started/reference/calculate-monitoring-consumption/

Regards,

Babar

[Julius_Loman · Accepted solution]

Every device and every dimension (for example interface) counts. So of you have 2 metrics pushed for 50 devices, you are consuming 100 licenses. Also oneagent plugins, activegate plugins and jmx metrics also consume this custom metric license.

[skrystosik]

So As I understand if I have 1 custom device with 10 metrics (2 dimensions per metric + 1 CUSTOM_DEVICE dimension), gives me 30 metrics consumed?

Sebastian

[Julius_Loman]

I think the answer is 21 in your case but I'm not sure what do you mean by one custom device dimension. On the case of custom metrics everything counts. That's why you can easily get over your limit when you have many dimensions and many instances.

[skrystosik]

When you create new custom metrics you can add own dimension but always there is one default that is named CUSTOM_DEVICE, which distinguish values between each custom devoce

[larry_roberts]

We having been playing with this and honestly I am having a hard time finding the exact value of these time series metrics which are not exactly cheap when you really can't do much of anything from the GUI with them.

We have been working on an ActiveGate plugin for Tibco EMS which runs on Solaris and also have been trying to building monitoring around JMS queues by using the JMX Editor, but we have found that the only way to set thresholds is from raw extensive code, there is no option to set custom alerts around them, and the AI does nothing with these metrics. With the JMX Editor you can't even make changes once a plugin is made - There is no edit options.

Other than a graph, I do not see what other value it brings in its current form.

Don't get me wrong, Dynatrace shines in a ton of stuff, but honestly this one has me scratching my head on where exactly the value is.

[skrystosik · Accepted solution]

Actually I've figured it out and working fine for me. I see all measures I see buy about calculation:

• Single measure with one dimension value counts as one
• Single measure with multiple dimensions counts as combination of distinct values of those dimensions
  
  
  • First dimension always is CUSTOM_DEVICE so if you have more than one this counts as number of custom devices
  • If you add extra dimensions this single measure will counts as combination of custom devices and distinct values of extra dimension
  • etc

It's quite simple to use couple of hundreds measures if you push to much dimensions 🙂

Sebastian

[larry_roberts]

It's quite simple to use couple of hundreds measures if you push to much dimensions 🙂

Yep! That is what we have found as well. And with all the limitations of what you can even do with those custom metrics as Dynatrace is today, we are considering picking another solution for that type of thing and sticking to Dynatrace for just things that are capable of coming from the OneAgent.

[Julius_Loman]

Agreed... Until you can do same things as with builtin metrics and create custom alerts the usability is quite limited.