Skip to content

[Cherry-Pick][Metrics] Support cpu-cache-block-num (#6390) #6391

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

Open
Copilot wants to merge 2 commits into
base: release/2.4
Choose a base branch
from
+13 −2
Open

[Cherry-Pick][Metrics] Support cpu-cache-block-num (#6390) #6391

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
93045bb
Initial plan
Copilot Feb 7, 2026
46abe80
[Metrics] Support cpu-cache-block-num
Feb 7, 2026
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
1 change: 1 addition & 0 deletions docs/online_serving/metrics.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,7 @@ After FastDeploy is launched, it supports continuous monitoring of the FastDeplo
| KV Cache | `fastdeploy:available_gpu_block_num` | Gauge | Available GPU blocks in cache (including unreleased prefix blocks) | count |
| KV Cache | `fastdeploy:free_gpu_block_num` | Gauge | Number of free GPU blocks in cache | count |
| KV Cache | `fastdeploy:max_gpu_block_num` | Gauge | Total number of GPU blocks initialized at startup | count |
| KV Cache | `fastdeploy:max_cpu_block_num` | Gauge | Total number of CPU blocks initialized at startup | count |
| KV Cache | `fastdeploy:available_gpu_resource` | Gauge | Ratio of available GPU blocks to total GPU blocks | % |
| KV Cache | `fastdeploy:gpu_cache_usage_perc` | Gauge | GPU KV cache utilization | % |
| KV Cache | `fastdeploy:send_cache_failed_num` | Counter | Total number of cache send failures | count |
Expand Down
3 changes: 2 additions & 1 deletion docs/zh/online_serving/metrics.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,8 @@
| KV缓存 | `fastdeploy:prefix_cpu_cache_token_num` | Counter | 位于 GPU 上的前缀缓存 token 总数 | 个 |
| KV缓存 | `fastdeploy:available_gpu_block_num` | Gauge | 缓存中可用的 GPU 块数量（包含尚未正式释放的前缀缓存块）| 个 |
| KV缓存 | `fastdeploy:free_gpu_block_num` | Gauge | 缓存中的可用块数 | 个 |
| KV缓存 | `fastdeploy:max_gpu_block_num` | Gauge | 服务启动时确定的总块数 | 个 |
| KV缓存 | `fastdeploy:max_gpu_block_num` | Gauge | 服务启动时确定的 GPU 总块数 | 个 |
| KV缓存 | `fastdeploy:max_cpu_block_num` | Gauge | 服务启动时确定的 CPU 总块数 | 个 |
| KV缓存 | `fastdeploy:available_gpu_resource` | Gauge | 可用块占比，即可用 GPU 块数量 / 最大GPU块数量| 百分比 |
| KV缓存 | `fastdeploy:gpu_cache_usage_perc` | Gauge | GPU 上的 KV 缓存使用率 | 百分比 |
| KV缓存 | `fastdeploy:send_cache_failed_num` | Counter | 发送缓存失败的总次数 | 个 |
Expand Down
2 changes: 2 additions & 0 deletions fastdeploy/cache_manager/prefix_cache_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -120,6 +120,7 @@ def __init__(
)

main_process_metrics.max_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.max_cpu_block_num.set(self.num_cpu_blocks)
main_process_metrics.available_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.free_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.available_gpu_resource.set(1.0)
Expand Down Expand Up @@ -428,6 +429,7 @@ def update_cache_config(self, cache_config):
self.node_id_pool = list(range(self.num_gpu_blocks + self.num_cpu_blocks))

main_process_metrics.max_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.max_cpu_block_num.set(self.num_cpu_blocks)
main_process_metrics.available_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.free_gpu_block_num.set(self.num_gpu_blocks)
main_process_metrics.available_gpu_resource.set(1.0)
Expand Down
9 changes: 8 additions & 1 deletion fastdeploy/metrics/metrics.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -158,6 +158,7 @@ class MetricsManager:
available_gpu_block_num: "Gauge"
free_gpu_block_num: "Gauge"
max_gpu_block_num: "Gauge"
max_cpu_block_num: "Gauge"
available_gpu_resource: "Gauge"
requests_number: "Counter"
send_cache_failed_num: "Counter"
Expand Down Expand Up @@ -235,7 +236,13 @@ class MetricsManager:
"max_gpu_block_num": {
"type": Gauge,
"name": "fastdeploy:max_gpu_block_num",
"description": "Number of total blocks determined when service started",
"description": "Number of total GPU blocks determined when service started",
"kwargs": {},
},
"max_cpu_block_num": {
"type": Gauge,
"name": "fastdeploy:max_cpu_block_num",
"description": "Number of total CPU blocks determined when service started",
"kwargs": {},
},
"available_gpu_resource": {
Expand Down
Loading