Skip to content

Update model server protocol with prefix cache reuse #1077

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jun 27, 2025
Merged

Update model server protocol with prefix cache reuse #1077

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 11 additions & 4 deletions docs/proposals/003-model-server-protocol/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,10 +21,10 @@ effort.
The corresponding metrics in vLLM are also shown in the table below, as vLLM is already integrated
into the reference endpoint picker implementation.

| Metric | Type | Description | vLLM metric |
| ----- | ---- | ---- | ---- |
| TotalQueuedRequests | Gauge | The current total number of requests in the queue.| `vllm:num_requests_waiting`|
| KVCacheUtilization| Gauge | The current KV cache utilization in percentage.| `vllm:gpu_cache_usage_perc`|
| Metric | Type | Description | vLLM metric | Triton TensorRT-LLM|
| ----- | ---- | ---- | ---- | ---- |
| TotalQueuedRequests | Gauge | The current total number of requests in the queue.| `vllm:num_requests_waiting`| `nv_trt_llm_request_metrics{request_type=waiting}`|
| KVCacheUtilization| Gauge | The current KV cache utilization in percentage.| `vllm:gpu_cache_usage_perc`| `nv_trt_llm_kv_cache_block_metrics{kv_cache_block_type=fraction}`|


### LoRA Adapter Serving
Expand All @@ -48,3 +48,10 @@ The model server MUST expose the following LoRA adapter metrics via the same Pro
* `running_lora_adapters`: A comma separated list of adapters that are currently loaded in GPU
memory and ready to serve requests. Example: `"running_lora_adapters": "adapter1, adapter2"`
* `waiting_lora_adapters`: A comma separated list of adapters that are waiting to be served. Example: `"waiting_lora_adapters": "adapter1, adapter2"`

### Prefix Cache Reuse

Starting from [v0.4.0](https://github.com/kubernetes-sigs/gateway-api-inference-extension/releases/tag/v0.4.0),
the EPP supports [prefix cache optimized request scheduling](https://gateway-api-inference-extension.sigs.k8s.io/guides/epp-configuration/prefix-aware/).
To benefit from the optimal prefix aware request scheduling, model servers SHOULD support prefix
cache reuse, such as the [vllm automatic prefix caching](https://docs.vllm.ai/en/latest/features/automatic_prefix_caching.html) feature.