Set policies for tables and files
Configure retention on tables and custom buckets. Policy changes apply to existing data, while new records and uploads inherit the active policy.
ART Datalake / Retention & management
Keep what matters. Put a boundary around what stays.
Persistent data needs an owner and a lifetime. Apply lifecycle rules where they belong, review changes, and understand storage usage as your workflows grow.
Retention & management in practice
Built for the way your data works.
Respect the purpose of each resource
Manage agent-memory retention through the agent, with pinning for context that must stay. Protected system resources use product-managed policies rather than ordinary bucket settings.
Manage with visibility
Scope resources to projects and environments. Review audit history by resource, action, actor, and time range, and monitor storage utilization and resource counts.
Illustrative workflow
Keep operational context without keeping everything forever
Choose appropriate lifetimes for an onboarding table and its document bucket. Retain selected agent context for follow-up work, then review the audit history and storage usage as the process runs.
- Choose a resource policy
- Apply it to the lifecycle
- Review usage and changes
Understand the lifecycle.
Tables and custom buckets default to indefinite retention. Expired data stops appearing in reads before background cleanup physically removes it. Some relationship constraints prevent automatic table retention.
A few useful details.
Is expiry the same as immediate physical deletion?
No. Expired data is excluded from reads immediately; a bounded lifecycle worker subsequently removes the stored data.
Is one retention policy applied to everything?
No. Tables, custom buckets, agent memories, and protected system resources each have their own lifecycle controls.
Start with your workflow
What should your data make possible?
Bring the records, files, and context behind your process. We’ll explore how Datalake fits.