Continuous Deployment for a Billion Dollar Order System
summary
Atlassian’s legacy ordering platform that has routed roughly a billion-dollar order system is converted from weekly dumps to per-merge releases, with every commit triggering a full test suite per commit and merges gated by pull requests. Deployments use Bamboo + Bitbucket + Ansible for staged rolling upgrades, SSH-based orchestration, and ticket-linked traceability via JIRA. Outcome: multiple safe deploys per day, much smaller blast radius for failures, automated rollback/monitoring hooks, and ongoing work to peel off microservices.
AI-written summary · about
chapters (14)
- 0:00Start
- 20:53Track your requests
- 22:05Work on a branch
- 23:28Continuously test
- 24:26Pull Request
- 26:09Merge and Release (staging)
- 27:02Promote to production
- 32:17Orchestration Agents
- 32:36Common Agents
- 34:09SSH Automation
- 36:05Our Solution
- 39:14Continuous Downtime?
- 40:55Deployment Environments
- 42:01Ticket-based management
related talks
34mEnabling a Continuous Deployment Mindset
26mDevOps and Lean in Legacy Environments
24mContinuous Deployment at Scale
43mContinuously deploying a distributed monolith
01h 01mA team's transition to Continuous Delivery
46mContinuous Deployment Workflows
34mContinuous Deployment for webapps based on Django
49mThe road to continuous deployment: a case study