Model the work: Real Life workloads with Ash & Oban
summary
Background job logic becomes brittle when jobs are treated as the source of truth, so this shows modeling work as mutable domain state using Ash resources, scheduled by AshOban/Ash Open triggers and guarded by AshStateMachine for workflows like welcome emails, webhooks and order fulfillment. The system runs reliably at over a million jobs a day, reconstructs missing jobs from data, preserves actor context via an actor persistor, and uses atomic updates and trigger-based scheduling to make retries and surgery predictable.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 12:43Modeling work with Ash & Oban
- 18:14Job records as god objects
- 22:47Enriching domain models
- 27:58State machines in workflows
- 33:16Advanced Ash Oban patterns
- 39:13Q&A
related talks
27mAsh Framework: The Power of Declarative Design
26mScaling Oban Applications
43mUnlocking the Potential of OTP: Proven Strategies for Application Dev
37mUsing Elixir for scheduling cloudnative DAG workflows
25mMigrating from C# to Elixir
58mModelling complex business domains with events
40mTurn Hours into Seconds- Concurrent Event Processing
01h 03mErlang & Akka Actors - A story of tradeoffs