Administration

Operating inference-cache at scale — observability, sizing, security, and troubleshooting.

This section covers the operational concerns of running inference-cache in production.

Monitoring & operations

Observability & Alerts

The opt-in Prometheus alert bundle, what each alert means and when it fires, and the scraping you must configure for the server, controller, and injected LMCache MP sidecars.

Index sizing

How the in-memory index consumes memory, the pod-budget table, the sizing formula, and the levers (TTL, eviction, quota) for keeping it in bounds.

Security

gRPC TLS

The locked TLS design, why gRPC is plaintext by default, and how to enable the opt-in TLS overlay.

Troubleshooting

Troubleshooting

The CacheBackend readiness runbook (per condition and reason) and the inferencecache doctor check catalog.

Last modified August 13, 2026: LMCache MP mode (#187) (40a9806)