Establish what is happening
Capture the incomplete result
Record endpoint, request identity, content type, size metadata, received item identifiers, and parser outcome for one case. Redact payload values. Note where the body ended and whether the client considered the result complete.
Compare declared and received data
Check collection counts, continuation metadata, content length, nested required fields, and stream termination. Compare the same fixture when complete. A parseable fragment is not proof that the server intended partial data.
Bound affected reads
Group cases by response size, endpoint, client revision, transport path, and dataset. Separate wire truncation, decoder stop, server limit, and valid partial response. Preserve any local record overwritten by incomplete data for review.
What to carry forward
Triage ends with one incomplete response, its first missing boundary, affected read cohort, and evidence gap list. Mark the result incomplete before changing storage. Route transport, parser, and contract questions separately. Record whether the incomplete response replaced a previously complete snapshot, and identify the consumer that accepted the shortened result.
Technical background: RFC Editor reference.
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