Skip to content

v0.7.18

Choose a tag to compare

View all tags
@SilenNaihin SilenNaihin released this 01 Sep 17:10
· 50 commits to main since this release
v0.7.18
f73da75
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Per-model failover policy (maximize_cache vs maximize_availability).

  • ExactModelPool.failover_mode + ExecutionSnapshot.failover_mode; next_route_candidate honors it.
  • maximize_cache: a THROTTLED (429) surfaces to the caller (no cold failover) to preserve prompt-cache affinity; operational deadness (auth/not-found/5xx/transport) and stalled-lane timeouts still fail over in both modes; genuine retryable 408 still redials the warm rung.
  • Authored GatewayPoolRecord.failover_mode carried through normalize_gateway_catalog.
  • Native crate unchanged (0.3.13).

(#705)

Assets 2
Loading