phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-04 04:04:14 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: api level request metrics via middleware

Browse source 
add RequestMetricsMiddleware which tracks key metrics related to each request the LLS server will recieve:

1. llama_stack_requests_total: tracks the total amount of requests the server has processed
2. llama_stack_request_duration_seconds: tracks the duration of each request
3. llama_stack_concurrent_requests: tracks concurrently processed requests by the server

The usage of a middleware allows this to be done on the server level without having to add custom handling to each router like the inference router has today for its API specific metrics.

Also, add some unit tests for this functionality

resolves #2597

Signed-off-by: Charlie Doern <cdoern@redhat.com>
This commit is contained in:
Charlie Doern 2025-07-11 20:52:32 -04:00
parent dbfc15123e
commit 49b729b30a
4 changed files with 433 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

3
docs/source/building_applications/telemetry.md
View file

@ -37,6 +37,9 @@ The following metrics are automatically generated for each inference request:
| `llama_stack_prompt_tokens_total` | Counter | `tokens` | Number of tokens in the input prompt | `model_id`, `provider_id` |
| `llama_stack_completion_tokens_total` | Counter | `tokens` | Number of tokens in the generated response | `model_id`, `provider_id` |
| `llama_stack_tokens_total` | Counter | `tokens` | Total tokens used (prompt + completion) | `model_id`, `provider_id` |
| `llama_stack_requests_total` | Counter | `requests` | Total number of requests | `api`, `status` |
| `llama_stack_request_duration_seconds` | Gauge | `seconds` | Request duration | `api`, `status` |
| `llama_stack_concurrent_requests` | Gauge | `requests` | Number of concurrent requests | `api` |
#### Metric Generation Flow