Using Dynatrace PurePath Data for Change Impact Analysis

[Vikash]

Hi

We use Dynatrace for monitoring our applications.  The infrastructure mapping and PurePath feature is great.  The problem detecting and business impact can be easily understood.

However what we would like to understand is the community's experience in using Dyntrace and the logged PurePath data to perform change impact analysis.  Let me explain what I mean.  

The business would like us to implement a feature.  We have done the technical analysis and identified the code/classes and components that need to change.  However as some of these classes and components as used by other features we would like to understand what features, UIs, upstream components and downstream components could be potentially impacted.

Can we use the logged PurePath and infrastructure discovery data to do this analysis? 

If we can, how would one approach this?

Does Dynatrace provide a UI/feature to do this?

Regards

Vikash

[Babar_Qayyum]

Hello @Vikash 

The best approach could be to verify everything in the UAT environment because as we know this is something like an assumption/or prediction sort of thing.
Also, you can compare service request performance and behavior over time once the changes have been done.

https://www.dynatrace.com/news/blog/compare-service-request-performance-behavior-time/

Regards,

Babar

[Radu · Accepted solution]

Hey Vikash,

Something similar to what you are describing can be done with Custom Services. An example here would be when trying to identify candidates for microservices to develop from an existing monolithic application. A detailed blog article (and great series to read overall) is available on the Dynatrace Blog.

Since you already know which classes you'll be working with, you can create a Custom Service by pointing Dynatrace to the entrypoint methods of that service. This helps you isolate that part of the code so that you can much more easily track its activity and understand its dependencies. You can use the Service Flow for downstream interactions (including DB activity) and the Backtrace for upstream interactions.

Best regards,

Radu

[Vikash]

Thanks for the replies - they have been very useful.

Regards

Vikash