Repository navigation

MAF-21239: feat(deploy): chart the request-observed KV cache hit per model - #189

Closed

hhk7734 wants to merge 6 commits into

mainfrom

MAF-21239-kv-cache-hit-panels

Closed

MAF-21239: feat(deploy): chart the request-observed KV cache hit per model#189
hhk7734 wants to merge 6 commits into
mainfrom
MAF-21239-kv-cache-hit-panels

Commits on Sep 29, 2026
- MAF-21239: feat(deploy): chart the request-observed KV cache hit per model
  
  hhk7734
  and
  claude
  committed
  8b0bc51
- MAF-21239: docs(website): explain the request-observed KV cache hit panels
  
  hhk7734
  and
  claude
  committed
  ff606ce
- MAF-21239: fix(deploy): say that client cancels skip the cache counters
  
  hhk7734
  and
  claude
  committed
  4f39442
- MAF-21239: fix(deploy): say that only vLLM needs the prompt tokens details flag
  
  hhk7734
  and
  claude
  committed
  d031902
- MAF-21239: feat(deploy): chart the KV cache hit from the engine token histograms
  
  hhk7734
  and
  claude
  committed
  9e84f77
- MAF-21239: docs(website): describe the histogram-based KV cache hit panels
  
  hhk7734
  and
  claude
  committed
  41081e9