Skip to content

TTL_MEMORY_TUNER.md

Latest commit

 

History

History
24 lines (16 loc) · 963 Bytes

TTL_MEMORY_TUNER.md

File metadata and controls

24 lines (16 loc) · 963 Bytes
Raw

TTL memory tuner (bridge)

Goal: Adjust ttl_seconds on opt-in workflow rows (tags include ttl_managed) based on bridge logic — shorter TTL for cold rows, extend hot ones when metadata supports it.

Entry points

  • Bridge verb: ttl_memory_tuner_run (see openclaw/bridge/bridge_token_budget_adapter.py).
  • Wrapper: openclaw/bridge/wrappers/ttl_memory_tuner.ainl
  • Runner: python3 openclaw/bridge/run_wrapper_ainl.py ttl-memory-tuner --dry-run

Environment

Variable Role
AINL_TTL_TUNER_TAG Required tag on memory metadata (default ttl_managed)

Safe rollout

  1. Always --dry-run first — no TTL writes.
  2. Tag a small set of non-critical records with ttl_managed.
  3. Run live on a narrow cron; monitor memory row counts and expiry behavior.
  4. Expand tags only after observation.

Wrong TTLs can expire data early — keep backups and test namespaces where possible.