Operating Cassandra in Kubernetes with Temporal
summary
Datadog runs Cassandra across hundreds of clusters and thousands of nodes on Kubernetes and replaces ad-hoc scripts with Temporal workflows that orchestrate init‑containers, StatefulSets, the Kubernetes API and long-running streaming repairs. Built components include a retryable health-check activity, batched node replacements in threes to preserve rack replica safety, pause/resume migration steps, integration with a DNS discovery service and distributed locks. The result automates multi-hour streaming operations, cuts manual coordination to near a single-button press, and materially reduces Ops time.
AI-written summary · about
chapters (11)
- 0:00Datadog and speakers intro
- 1:06What is Temporal
- 3:14Runbooks to workflows
- 6:45Temporal setup
- 7:49Health checks basics
- 11:43Migrations with Temporal
- 13:09Robust migration workflows
- 16:34Reusable workflows
- 17:57Single-button migration
- 19:21Temporal future
- 20:48Q&A
related talks
35mBuild fault tolerant, distributed apps with Kubernetes and Temporal
42mDistributed Data Stores on Kubernetes
26mRethinking Cloud-Native Patterns on Kubernetes
21mDeclarative & Imperative Kubernetes Operations with Temporal & M3
41mOrchestrating Data Pipelines at Scale with Temporal
30mOrchestrate Your Business Workflows with Temporal
44mKubernetes at Datadog Scale
33mLessons learned using Temporal to run Temporal Cloud