FROM THE BENCH
Longer pieces about the things that bite in production — the ones worth writing down properly rather than leaving in a commit message.
Most distributed-system incidents aren't caused by hard failures. They're caused by transient ones handled badly.
429 · 503 · backoff · jitter · idempotency · Tenacity
When concurrent requests race to insert the same record, most teams reach for retry logic. Django's get_or_create handles it in one line.
IntegrityError · savepoint · unique index · upsert · PostgreSQL