Test, run history, and the diagnostic loop
1. Test, run history, and the diagnostic loop
Building a flow is half the job. Knowing why a run failed — or why a "green" run did the wrong thing — is the other half. This video covers the Test pane, run history, and the diagnostic loop that ties them together.2. The Test pane
The Test pane is how you trigger a run on demand. Open the flow in the designer, click Test, and pick one of two modes. Manually fires the trigger immediately using the inputs you supply, which is fastest, but the trigger payload is fabricated. Automatically waits for the next real trigger event, which is what you want when the trigger itself is what you're debugging. Either way, the test run starts and the designer switches into a live view where you can watch each action complete in real time. Green ticks for success, red X for failure, every step time-stamped.3. Run history
Run history is the system of record for everything the flow has ever done, at least for the last 28 days, which is the standard retention window. Every run shows start time, duration, status, and the trigger that fired it. Filter by status to find the red ones, by date for an incident window, by trigger to isolate one source. Click any row and you drop into the run details, with every action expanded in order, with inputs and outputs on each, exactly as it executed. This is the panel you live in when something went wrong.4. Inputs and outputs: the diagnostic superpower
Inside a run, every action exposes its inputs, what was passed in, and its outputs, what came back. Click an action to expand it, then click Show raw outputs and the panel switches to the actual JSON payload the connector returned. This is your diagnostic superpower. SharePoint columns have a display name in the UI, Contract Value, and an internal name in the JSON, `ContractValue`. Dynamic content uses the internal names, so a flow that looks right in the designer can quietly be referencing the wrong column. Raw outputs is where you see the truth.5. The diagnostic loop
Diagnosis follows a loop. Read the run history. Identify the failed step, the first red one in the chain. Expand it and read the error output; connector errors carry both a human-readable message and a structured error code. Compare the inputs the step received to what you expected. That's where most bugs live, in a dynamic reference that resolved to something different than you thought. Make the fix in the designer, save the flow, then Resubmit the same run from history to replay it with the original trigger payload. Loop until green.6. Flow Checker as the pre-flight check
Flow Checker is the static-analysis tool, available in the designer toolbar. It runs before you save and surfaces findings the flow has, such as missing required fields, undefined variables, actions with no error handling, and defaults that might trip you up later. The catch is that it mixes hard blockers with stylistic warnings in the same list. An undefined variable will fail at runtime. Don't auto-fix everything Flow Checker raises; some of its suggestions are wrong for your flow.7. Let's practice!
Time to find and fix a deliberately seeded bug in the renewal flow.Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.