How Canva Built Simple, Auditable, and Maintainable Data Retention
summary
Canva re-architects deletion with a central deletion service using a star topology that replaces brittle event chains, generates per-service wiring from a small parameter file so each service needs only a few lines of code, and executes queued deletes and replayable audits. Production data snapshots are taken from every S3, DynamoDB and RDS instance in a locked mirror to delta against 16 activity scripts, run weekly across ~3,000 persistent stores, and can time-travel simulate a 7‑year retention for evidence.
AI-written summary · about
chapters (7)
- 0:00Data retention at Canva
- 1:14Canva’s retention model
- 4:23Cascading deletion failures
- 6:26Star topology deletion
- 9:56Implementing deletion model
- 11:42Auditing data deletion
- 14:50Key takeaways
related talks
16mDeleting Data at Organizational Scale
24mCheck before You Change: Preventing Correlated Failures in Service Updates
35mIntelligent Deployments at Scale: Effortless Continuous Deploy...
26mCI/CD For Data – Building Data Development Environments With Open Source Stacks
25mContent creation under a new creative paradigm from on-premises to cloud
51mCutting edge Continuous Delivery: Automated Canary Analysis through Spring based Spinnaker
23mHow Atlassian is migrating 40PB data with 99.99% reliability
25mDelete This: Decommissioning Servers at Scale