Advice Needed: Best Practices for Service Detection Rules

[lwaldrop]

Update: I enabled a full web service rule to detect a service as a web request service by setting a condition on <NameService.svc>. It seems to be tracking correctly, so I’m considering making this the default rule without conditions, as it applies to multiple scenarios. Has anyone experienced a service being split into two entities: a web request service and a web service?

######

Hi All,

I'm seeking guidance on best practices for creating a service detection rule. We have a service in our environment that is being detected as two separate entities, which is causing discrepancies in our alerting. Currently, it's recognized as <NameEntryPoint.API> and <NameService.svc>, but it should be detected as <NameEntryPoint.API>\<NameService.svc>.

I've reviewed the documentation (Service detection rules - Dynatrace Docs), but I'm looking for specific best practices or recommendations on how to address this issue effectively. Any insights or advice would be greatly appreciated.

Sincerely,

Lance

[GosiaMurawska]

Hi @lwaldrop! How did it go, did you apply it as the default rule? It would be great if you share your case with the Community 🙂 

[lwaldrop]

Hi @GosiaMurawska. No I haven't tried this yet as other things come up. I was also hoping someone else may have tried it first. 🙂

[Peter_Youssef]

Hi @lwaldrop 

Simply it depends on the actual detected request and service. the documentation is giving a proper guidance for:

Internally through the monitored servers

• Web request.
• Web Service.

Externally:

• Web request.
• Web Service.

For example

• Regarding the external APIs or External endpoints you can configure the public domain name.

For internally monitored services, it depends on mainly how many segments the service is composed of.

Tip:

• To facilitate the composition of service detection rules, have a deeper look into the services properties && Tags section as it contains the most unique attributes based on which you can draft the service detection rules that fits the current setup.
• You can try multiple attempts on testing environments and you will notice how it works practically and then take notes of the actual findings for further enhancements.

Hoping it adds value.

BR,

Peter