Keep the lesson for the next incident
Document the trigger and measured boundary
Record the exact input sequence, route, dataset, viewport, and device class. Describe the main-thread work observed, the first responsive signal, and the behavior that made the delay feel frozen. Link baseline and fixed traces, and label any stack attribution that remains uncertain.
Set an interaction-focused guard
Add a repeatable journey or review check that measures response to the primary input with representative data. Define the observation as input feedback, cancellation, or usable output rather than an arbitrary task count. Assign ownership for interpreting a breach and checking whether the journey still reflects usage.
Revisit when data shape changes
Tie a follow-up review to list growth, a new transformation, or a rendering-library change. Re-run the interaction after those changes and compare task distribution, not only the total. If the workflow gains a new branch, add a separate case instead of assuming one trace covers it.
What to carry forward
The learning record should preserve what froze, why the evidence pointed to a main-thread boundary, how responsiveness was restored, and which guard remains. Future engineers should be able to reproduce the interaction before editing and recognize when a new data shape invalidates the old evidence.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow