Overview: Most Canvas apps reach back to 3rd parties or internal storage e.g. SQL, Dataverse, blobs to persist data. In an ideal world we want to be able to identify error and performance concerns and be able to trace quickly.
Scenario: A canvas app performs a search against a custom connector, that goes to an external API, for example search's for Company Numbers from an Open API.
- The annotated diagram, in orange records the button pressed to run the search. This only shows if the "Upcoming feature", "Enable Azure Application Insights correlation tracing" is enabled.
- A custom connector, in blue, makes a call to an REST/Open API.
- The last call is to an external API, highlighted in purple, (in this case it was actually a mocked APIM endpoint, but the icon would change it it was totally external such as call API search to the IRS, HMRC, Inland Revenue.
Tip: Always create App Insight Instances using a workspace and not the classic-mode app insights as it is being deprecated by Microsoft in early 2024.
App Insights Understanding:
App Insights setup using the Workspace-based setup (Log Analytics) can be queried in two ways:
- Query thru App Insights
- Query thru the Log Analytics workspace (the Kusto/KQL is slightly different, but it's rather minor)
Tip: Power Automate has a connector to Log Analytics so it's good to use this for flows so you can trace canvas apps using flow journeys. Most people tend to build a custom connector to a function that uses the AppInsights SDK. I've used both and they are both valid approaches and shown in the annotated diagram below.
 |
| The two options for logging Flows into App Insights. |
Note: If you create a new App Insights workspace-based instance remember to update the loggers in all you Azure Services to the new instance (app key). For example functions, APIM and Service Bus are common components.
Note: You can log to multiple workspace/app insights in a tenant and the correlations will be retrieved so you see the full history, assuming you have permissions to all log sources.
Learning: the instrumentation key for app insights has 3 parts to it:
1) instrumentation key (basically a unique identified to find and allow logs to be saved into AppInsights,
2) ingestion endpoint (URL for the log), and
3) monitoring metric endpoint (URL for metrics/performance counters/live metrics/failed requests/).
Here is an example and you can see the 3 parts:
InstrumentationKey=2675bxxx-xxxb-xxxx-bf5558009ccf;IngestionEndpoint=https://uksouth-1.in.applicationinsights.azure.com/;LiveEndpoint=https://uksouth.livediagnostics.monitor.azure.com/
