Distributed tracing across multiple services in Power Platform
A useful resource for architects and developers - Microsoft Power Platform and Copilot Studio Architecture Center - Power Platform | Microsoft Learn
It goes into some depth on areas you probably thought you might already understand
it discussed data auditing but then discusses how you can audit business events. You don't get this with standard auditing
Distributing tracing across multiple services in power platform
This section I found very interesting and its the most comprehensive documentation I have seen on the topic
Distributed tracing across multiple services in Power Platform - Power Platform | Microsoft Learn
The diagram below sums up the different services that can now push data to application insights, which lets you monitor and alert on the log files.
Microsoft like to share services between services rather than recreating a version of the service for all their different applications.
Entra ID - authentication
Power BI - Reporting
Exchange - Email
Application insights - monitoring and alerting
This is a good approach because it creates one good service, rather than many half good services. The down
Discusses how you can trace across different services and how they work. It talks about the key monitoring tools like Azure monitor and application insights, with some tips on writing KQL.
if you use Managed environments, most of the services can put telemetry (logs) into application insights.
Implementing tracing
tracing isn't a topic covered often, the page above is excellent. It explains how it works, potential use cases. It discusses W3C Trace Context.
I haven't heard of the W3C trace context standard, which according to the document has these benefits
End-to-end visibility: Trace context provides end-to-end visibility into a transaction, helping you understand how requests propagate through your system.
Performance monitoring: It lets you monitor the performance of individual services and identify bottlenecks.
Error diagnosis: It helps you diagnose errors by tracing the path of a request and identifying where failures occur.
Dependency tracking: It allows you to track dependencies between services and understand how they interact.
It goes onto to discuss how you can add tracing to plugins, Dataverse integration and other scenarios which are beyond the basics of tracing.
Very good read, I would recommend it.