Cadence: The New Open Source Project for Building Complex...
summary
Cadence is an open source orchestration engine that preserves workflow state durably so long running business processes can be replayed and recovered. It runs as a single binary (four roles can be split later), uses pluggable storage and visibility backends like Cassandra MySQL Postgres Dynamo Mongo and Kafka for replication, exposes workflows, activities, signals and timers, and ships a web UI plus default metrics. Uber runs 12 billion workflows a month with about 300 billion updates monthly and three nines availability. V1 is released and V2 redesign is planned.
AI-written summary · about
chapters (10)
- 0:00Start
- 0:34What is a Workflow?
- 4:08Non-Functional Requirements
- 6:34Stateless Approach
- 13:59Development (deep-dive)
- 19:44Key Concepts
- 22:22Metrics
- 23:31Cadence @Uber
- 24:34V1 Release!
- 29:12References
related talks
32mScalable big data processing pipeline with Cadence
25mSpinning Your Drones with Cadence Workflows and Apache Kafka
17mRapidly Build Distributed Applications using Orchestration
40mSupercharging Cloud Orchestration with Temporal Open Source
44mGetting started with Orchestration and Choreography in Distributed Applications
50mCadre: Building Reliable Distributed Stateful Applications
42mOrchestration vs Choreography, A Guide To Composing Your Monolith
49mThe Benefits of Orchestration in Distributed Systems