Ensuring Reliable Distributed Transactions in Logistics
summary
Defines four concrete tenets that make distributed logistics workflows robust: replayability so producers can resend events from hours or weeks, guaranteed idempotency via unique message IDs, clear separation of event vs command flows, and explicit schema ownership. Shows an outbox pattern for replay, gives numeric heuristics used by the tracking engine (ignore ETA jumps of ~10 or 21 days, flag >5 day changes after departure), routes suspicious events to Slack for ops approval, and replaces manual gating with a retrainable ML classifier.
AI-written summary · about
chapters (4)
related talks
56mFrom Unreliable to Resilient: Mastering Distributed Systems
57mSafe Passage: Messaging in a distributed systems environment
45mDo Distributed Systems they said. It'll be fun they said.
51mComplex Event Flows in Distributed Systems
39mTrust Deterministic Execution to Scale & Simplify Your Systems
47mComplex event flows in distributed systems
01h 38mBuilding Reliable Distributed Systems
34mAtomicity and Consistency Across Databases and Messaging Systems