Topic
Publish/subscribe designs in which producers emit facts and consumers independently decide to react, including broker retention and replay behavior.
Ian Cooper says event-driven orgs of up to about five teams can manage async APIs on shared knowledge of each other's services. His model for doing it at scale begins with the address, binding and contract each endpoint has to describe to its consumers.
Reality
A developer's write-up of running the outbox pattern in production finds the textbook loop breaks on error to preserve ordering, and never counts attempts. That combination jams forever.
Reality
AWS launched an EventBridge custom bus that a whole organization shares through RAM, with 10,000 subscribers per bus by default. Platform teams can move onto it and retire the cross-account rules and bus-to-bus hops they wired by hand.
Reality
A dev.to post argues that processing.guarantee=exactly_once_v2 charges coordination overhead to every message and still leaves external side effects for you to deduplicate by hand. The five-line alternative it offers has a gap of its own.
Reality
Kafka promises at-least-once delivery, so a consumer that dies after the payment call but before the commit replays the event on restart. A dev.to tutorial puts the guard in a database primary key, keyed on an event ID the producer sets.
Reality
A dev.to teardown makes the case for event-triggered Node workers with a slow poll behind them, and the idle-cost numbers it publishes come to about 138,000 empty queries a day across eight instances.
Reality
An experiment written up on dev.to publishes outbox events from an in-memory queue of committed event IDs and keeps polling only as a recovery scan, so that scan's interval now sets worst-case delivery latency.
Reality
A dev.to spec defines a silent failure rate that platform run history cannot produce, computed by reconciling events a controller you own fires against receipts posted back by every action in the workflow. It is free under CC BY 4.0.
Reality
The Inbox pattern makes a consumer idempotent only when the dedup row and the business write commit together, and a Spring Boot walkthrough works through what each of the two wrong orderings costs you.
Reality
A multi-agent orchestrator held its cross-site correlation window in process memory. It fired live on the first try, for the wrong reason, and the fix was a shared record in Firestore.
Reality
A settlement engineer trades the transactional outbox for producer-side causality plus consumer-side dedup. The reasoning holds up, and the half that carries it is the half not shown.
Reality
A dev.to write-up argues that synchronous agent chains break on restarts, slow tools and approval waits, and that the fix is an event bus plus a policy layer before any real action.
Reality