Topic
Engineering Practice
30 articles about Engineering Practice, newest first.
Browse the complete article archive →A Flaky Test Passed on Retry: Diagnose It Before You Trust the Build
Investigate a passing retry with a real isolated test-order example, preserved failure logs, a fixture repair, and clear limits on what repeated passes prove.
A Test Passes Locally but Fails in CI: Reproduce the Environment Difference
Compare effective environments before changing a failing CI test. A real local directory fixture preserves the error and checks a repair in both conditions.
Add a Regression Test That Fails for the Original Bug
Write a regression test that fails on the original bug, then verify the unchanged expectation and adjacent boundaries with an actual Python example.
An Alert Keeps Waking the Team: Check Whether It Requires Human Action
Review a noisy alert with a completed response map and synthetic rule replay. Keep urgent capacity and coverage risks visible before changing paging routes.
An API Rate-Limits Your Integration: Budget Requests Across Workers
Budget API requests across workers using one queue, reserve and pause state. Inspect a complete offline recovery trace before changing a real integration.
An Application Rollback Does Not Undo a Database Change: Plan Both Paths
Plan code rollback and database recovery separately with a tested SQLite compatibility matrix and a fictional decision record that preserves data limits.
API Pagination on Changing Data: Reduce Missing and Duplicate Results
Compare offset and cursor pagination on changing records, then test ordering, snapshot, retry and deduplication limits before claiming a complete API scan.
Before Shipping a New Integration, Test Timeouts, Retries, and Duplicate Requests
Test response loss after commit, stable idempotency keys, conflicting replays, and retry deadlines with a complete local integration example and honest limits.
Bug Severity vs. Priority: Run Triage with Impact and Reproduction Evidence
Separate bug impact from repair order with six worked triage decisions, a bounded reproduction, and a high-severity case whose priority changes with exposure.
Code Coverage Went Up: Check Which Behaviors Are Still Untested
Compare executed statements, visited branches, and useful assertions with a complete Python example where full coverage still misses three deliberate mistakes.
Create an SBOM That Can Answer Which Release Contains a Dependency
Trace a dependency to a release with an SBOM, artifact digests and a worked graph query. Keep missing inventory and vulnerability assessment separate.
Deprecate an API Without Surprising Its Consumers
Deprecate an API with a consumer ledger, precise notice and header dates, and a retirement review that keeps inactive-looking and unknown callers visible.











