Metrics

Overview

Dynamo provides built-in metrics capabilities through the Dynamo metrics API, which is automatically available whenever you use the DistributedRuntime framework. This document serves as a reference for all available metrics in Dynamo.

For visualization setup instructions, see the Prometheus and Grafana Setup Guide.

For creating custom metrics, see the Metrics Developer Guide.

Environment Variables

Getting Started Quickly

This is a single machine example.

Start Observability Stack

For visualizing metrics with Prometheus and Grafana, start the observability stack. See Observability Getting Started for instructions.

Launch Dynamo Components

Launch a frontend and vLLM backend to test metrics:

$
# Start frontend (default port 8000, override with --http-port or DYN_HTTP_PORT env var)
$
$ python -m dynamo.frontend
$
$
# Enable backend worker's system metrics on port 8081
>
$ DYN_SYSTEM_PORT=8081 python -m dynamo.vllm --model Qwen/Qwen3-0.6B  \
>
   --enforce-eager --no-enable-prefix-caching --max-num-seqs 3

Wait for the vLLM worker to start, then send requests and check metrics:

$
# Send a request
$
curl -H 'Content-Type: application/json' \
>
-d '{
>
  "model": "Qwen/Qwen3-0.6B",
>
  "max_completion_tokens": 100,
>
  "messages": [{"role": "user", "content": "Hello"}]
>
}' \
>
http://localhost:8000/v1/chat/completions
$
$
# Check metrics from the backend worker
$
curl -s localhost:8081/metrics | grep dynamo_component

Exposed Metrics

Dynamo exposes metrics in Prometheus Exposition Format text at the /metrics HTTP endpoint. All Dynamo-generated metrics use the dynamo_* prefix and include labels (dynamo_namespace, dynamo_component, dynamo_endpoint) to identify the source component.

Example Prometheus Exposition Format text:

# HELP dynamo_component_requests_total Total requests processed
# TYPE dynamo_component_requests_total counter
dynamo_component_requests_total{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate"} 42
# HELP dynamo_component_request_duration_seconds Request processing time
# TYPE dynamo_component_request_duration_seconds histogram
dynamo_component_request_duration_seconds_bucket{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate",le="0.005"} 10
dynamo_component_request_duration_seconds_bucket{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate",le="0.01"} 15
dynamo_component_request_duration_seconds_bucket{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate",le="+Inf"} 42
dynamo_component_request_duration_seconds_sum{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate"} 2.5
dynamo_component_request_duration_seconds_count{dynamo_namespace="default",dynamo_component="worker",dynamo_endpoint="generate"} 42

Metric Categories

Dynamo exposes several categories of metrics:

• Frontend Metrics (dynamo_frontend_*) - Request handling, token processing, and latency measurements
• Component Metrics (dynamo_component_*) - Request counts, processing times, byte transfers, and system uptime
• Specialized Component Metrics (e.g., dynamo_preprocessor_*) - Component-specific metrics
• Engine Metrics (Pass-through) - Backend engines expose their own metrics: vLLM (vllm:*), SGLang (sglang:*), TensorRT-LLM (trtllm_*)

Runtime Hierarchy

The Dynamo metrics API is available on DistributedRuntime, Namespace, Component, and Endpoint, providing a hierarchical approach to metric collection that matches Dynamo’s distributed architecture:

• DistributedRuntime: Global metrics across the entire runtime
• Namespace: Metrics scoped to a specific dynamo_namespace
• Component: Metrics for a specific dynamo_component within a namespace
• Endpoint: Metrics for individual dynamo_endpoint within a component

This hierarchical structure allows you to create metrics at the appropriate level of granularity for your monitoring needs.

Available Metrics

Note: Labeled metrics (HistogramVec, CounterVec, GaugeVec) register a metric family, not individual time series. A series for a given label combination only appears at /metrics after the first with_label_values(...) call for that combination — i.e., after the first matching request is served. For example, dynamo_frontend_request_duration_seconds{model="Qwen/Qwen3-0.6B"} will not appear on a freshly-started frontend until a request for that model is handled. This is expected Prometheus client behavior, not a missing metric.

Backend Component Metrics

Backend workers (python -m dynamo.vllm, python -m dynamo.sglang, etc.) expose dynamo_component_* metrics on the system status port (configurable via DYN_SYSTEM_PORT, disabled by default). In Kubernetes the operator typically sets DYN_SYSTEM_PORT=9090; for local development you must set it explicitly (e.g. DYN_SYSTEM_PORT=8081).

The core Dynamo backend system exposes metrics at the /metrics endpoint with the dynamo_component_* prefix for all components that use the DistributedRuntime framework:

• dynamo_component_inflight_requests: Requests currently being processed (gauge)
• dynamo_component_request_bytes_total: Total bytes received in requests (counter)
• dynamo_component_request_duration_seconds: Request processing time (histogram)
• dynamo_component_requests_total: Total requests processed (counter)
• dynamo_component_response_bytes_total: Total bytes sent in responses (counter)
• dynamo_component_uptime_seconds: DistributedRuntime uptime (gauge). Automatically updated before each Prometheus scrape on both the frontend (/metrics on port 8000) and the system status server (/metrics on DYN_SYSTEM_PORT when set).

Access backend component metrics:

$
# Set DYN_SYSTEM_PORT to enable the system status server
$
DYN_SYSTEM_PORT=8081 python -m dynamo.vllm --model <model>
$
curl http://localhost:8081/metrics

Specialized Component Metrics

Some components expose additional metrics specific to their functionality:

• dynamo_preprocessor_*: Metrics specific to preprocessor components

Frontend Metrics

Important: The frontend and backend workers are separate components that expose metrics on different ports. See Backend Component Metrics for backend metrics.

The Dynamo HTTP Frontend (python -m dynamo.frontend) exposes dynamo_frontend_* metrics on port 8000 by default (configurable via --http-port or DYN_HTTP_PORT) at the /metrics endpoint. Most metrics include model labels containing the model name:

• dynamo_frontend_active_requests: Number of requests currently being handled by the frontend, from HTTP handler entry until the response stream completes (gauge). This is the top-level in-flight count with no stage breakdown.
• dynamo_frontend_stage_requests: Number of requests currently in a given frontend pipeline stage (gauge, labels: stage, phase). See Stage and phase labels below.
• dynamo_frontend_inflight_requests: Inflight requests (gauge). Deprecated — kept for backward compatibility; prefer dynamo_frontend_active_requests, which has identical semantics with a clearer name.
• dynamo_frontend_queued_requests: Number of requests in HTTP processing queue (gauge). Deprecated — kept for backward compatibility; the “waiting for first token” window is now the sum of dynamo_frontend_stage_requests across the preprocess, route, and dispatch stages.
• dynamo_frontend_disconnected_clients: Number of disconnected clients (gauge)
• dynamo_frontend_input_sequence_tokens: Input sequence length (histogram)
• dynamo_frontend_cached_tokens: Number of cached tokens (prefix cache hits) per request (histogram)
• dynamo_frontend_inter_token_latency_seconds: Inter-token latency (histogram)
• dynamo_frontend_output_sequence_tokens: Output sequence length (histogram)
• dynamo_frontend_output_tokens_total: Total number of output tokens generated (counter)
• dynamo_frontend_request_duration_seconds: LLM request duration (histogram)
• dynamo_frontend_requests_total: Total LLM requests (counter)
• dynamo_frontend_time_to_first_token_seconds: Time to first token (histogram)
• dynamo_frontend_model_migration_total: Total number of request migrations due to worker unavailability (counter, labels: model, migration_type)

Access frontend metrics:

$
curl http://localhost:8000/metrics

Stage and phase labels

dynamo_frontend_stage_requests decomposes the lifetime of an active frontend request into three sequential pipeline stages. A request is counted in exactly one stage at a time (via an RAII guard that increments on stage entry and decrements on stage exit), and is counted in dynamo_frontend_active_requests for its entire lifetime. Between stages — and after dispatch exits while the backend is streaming tokens — the request is still in active_requests but in no stage_requests bucket.

stage label values:

phase label values:

Derived signals operators commonly want. These are cluster-wide totals across all frontend pods. stage_requests has no model label, so you cannot split these by model; add by (pod) or by (instance) to any sum(...) below if you need per-pod visibility. The stage filter stage=~"preprocess|route|dispatch" is used explicitly to keep the “pre-first-token” semantic stable if additional stages (e.g. postprocess) are added in the future.

• Requests waiting for a worker to start generating (the old “queued” semantic): sum(dynamo_frontend_stage_requests{stage=~"preprocess|route|dispatch"}) — i.e. still in preprocess, route, or dispatch.
• Requests currently being processed by a backend worker: use the worker-side gauge sum(dynamo_component_inflight_requests{dynamo_component="backend",dynamo_endpoint="generate"}) — this is the authoritative count, available whenever DYN_SYSTEM_PORT is set on workers (see Backend Component Metrics).
  • Frontend-perspective variant (useful if worker metrics aren’t being scraped, or when sizing frontend pods rather than workers): sum(dynamo_frontend_active_requests) - sum(dynamo_frontend_stage_requests{stage=~"preprocess|route|dispatch"}). This differs from the worker gauge because its window starts when the first token arrives at the frontend and extends through streaming to the client, so it includes transit and client-buffering time.
• Router saturation: sum(dynamo_frontend_stage_requests{stage="route"}) spiking indicates workers can’t be selected fast enough (e.g. all backends busy, KV-router queue full).
• Backend prefill saturation: sum(dynamo_frontend_stage_requests{stage="dispatch"}) spiking indicates the backend is slow to produce first tokens.

Deprecated frontend gauges

The following gauges are still emitted but will be removed in a future release. They were superseded by the gauges above as part of the frontend-metrics rework (PR #8162). Dashboards and alerts should migrate off them.

Model Configuration Metrics

The frontend also exposes model configuration metrics (on port 8000 /metrics endpoint) with the dynamo_frontend_model_* prefix. These metrics are populated from the worker backend registration service when workers register with the system. All model configuration metrics include a model label.

Runtime Config Metrics (from ModelRuntimeConfig): These metrics come from the runtime configuration provided by worker backends during registration.

• dynamo_frontend_model_total_kv_blocks: Total KV blocks available for a worker serving the model (gauge)
• dynamo_frontend_model_max_num_seqs: Maximum number of sequences for a worker serving the model (gauge)
• dynamo_frontend_model_max_num_batched_tokens: Maximum number of batched tokens for a worker serving the model (gauge)

MDC Metrics (from ModelDeploymentCard): These metrics come from the Model Deployment Card information provided by worker backends during registration. Note that when multiple worker instances register with the same model name, only the first instance’s configuration metrics (runtime config and MDC metrics) will be populated. Subsequent instances with duplicate model names will be skipped for configuration metric updates.

• dynamo_frontend_model_context_length: Maximum context length for a worker serving the model (gauge)
• dynamo_frontend_model_kv_cache_block_size: KV cache block size for a worker serving the model (gauge)
• dynamo_frontend_model_migration_limit: Request migration limit for a worker serving the model (gauge)

Request Processing Flow

Deprecated framing. The two-metric model below (inflight vs. HTTP queue) describes the legacy dynamo_frontend_inflight_requests and dynamo_frontend_queued_requests gauges and is kept only to help operators reading existing dashboards. New work should use dynamo_frontend_active_requests and the per-stage dynamo_frontend_stage_requests gauges described under Stage and phase labels.

This section explains the distinction between two key metrics used to track request processing:

1. Inflight: Tracks requests from HTTP handler start until the complete response is finished
2. HTTP Queue: Tracks requests from HTTP handler start until first token generation begins (including prefill time)

Example Request Flow:

curl -s localhost:8000/v1/completions -H "Content-Type: application/json" -d '{
  "model": "Qwen/Qwen3-0.6B",
  "prompt": "Hello let's talk about LLMs",
  "stream": false,
  "max_tokens": 1000
}'

Timeline:

Concurrency Example: Suppose the backend allows 3 concurrent requests and there are 10 clients continuously hitting the frontend:

• All 10 requests will be counted as inflight (from start until complete response)
• 7 requests will be in HTTP queue most of the time
• 3 requests will be actively processed (between first token and last token)

Key Differences:

• Inflight: Measures total request lifetime including processing time
• HTTP Queue: Measures queuing time before processing begins (including prefill time)
• HTTP Queue ≤ Inflight (HTTP queue is a subset of inflight time)

Router Metrics

The router exposes metrics for monitoring routing decisions and overhead. Defined in lib/llm/src/kv_router/metrics.rs.

For router deployment modes, see the Router Guide. For router flags and tuning, see Configuration and Tuning.

Metrics Availability by Configuration

Not all metrics appear in every deployment. The chart below shows which metric groups are registered and populated in each configuration:

Key:

• Registered and populated: Metric appears at /metrics with real values
• Registered, always zero: Metric appears at /metrics but the counter/histogram is never incremented (useful for dashboards that expect the metric to exist)
• Not registered / Not created: Metric does not appear at /metrics at all

Scrape endpoints:

• Frontend: /metrics on HTTP port (default 8000, configurable via --http-port or DYN_HTTP_PORT)
• Standalone router: /metrics on DYN_SYSTEM_PORT (must be set explicitly; default is -1 / disabled)
• Backend workers: /metrics on DYN_SYSTEM_PORT (separate from frontend metrics)

Router Request Metrics (dynamo_component_router_*)

Histograms and counters for aggregate request-level statistics. Eagerly registered via from_component() with the DRT MetricsRegistry hierarchy. On the frontend, exposed at /metrics on the HTTP port (default 8000) via the drt_metrics bridge. On the standalone router (python -m dynamo.router), exposed on DYN_SYSTEM_PORT when set. Populated per-request when --router-mode kv is active; registered with zero values in non-KV modes.

All metrics carry the standard hierarchy labels (dynamo_namespace, dynamo_component, dynamo_endpoint).

Per-Request Routing Overhead (dynamo_router_overhead_*)

Histograms (in milliseconds) tracking the time spent in each phase of the routing decision for every request. Registered on the frontend port (default 8000) at /metrics with a router_id label (the frontend’s discovery instance ID). These metrics are only created when the frontend has DRT discovery enabled (i.e., --router-mode kv); they do not appear in non-KV modes or on the standalone router.

Router Queue Metrics (dynamo_frontend_router_queue_*)

Gauge tracking the number of requests pending in the router’s scheduler queue. Only registered when --router-queue-threshold is set. Labeled by worker_type to distinguish prefill vs. decode queues in disaggregated mode.

Labels: worker_type (prefill or decode)

KV Indexer Metrics

Tracks KV cache events applied to the router’s radix tree index. Only appears when --router-kv-overlap-score-weight is greater than 0 (default) and workers are publishing KV events. Will not appear if --router-kv-overlap-score-weight 0 is set or no KV events have been received.

Additional labels: status (ok / parent_block_not_found / block_not_found / invalid_block), event_type (stored / removed / cleared)

Per-Worker Load and Timing Gauges (dynamo_frontend_worker_*)

These appear once workers register and begin serving requests. They are registered on the frontend’s local Prometheus registry (not component-scoped) and do not carry dynamo_namespace or dynamo_component labels. These metrics are frontend-only and are not available on the standalone router.

Labels:

In disaggregated mode, the worker_type label shows both "prefill" and "decode" values; in aggregated mode, all workers report as "decode".

NIXL Telemetry Metrics

NIXL exposes its own Prometheus metrics on a separate port from Dynamo metrics. These metrics track KV cache and embedding data transfers and are only populated during disaggregated serving or multimodal embedding transfers.

To enable, set these environment variables on your worker process:

$
# Prefill worker
$
NIXL_TELEMETRY_ENABLE=y NIXL_TELEMETRY_EXPORTER=prometheus \
>
  NIXL_TELEMETRY_PROMETHEUS_PORT=19090 DYN_SYSTEM_PORT=8081 \
>
  python -m dynamo.vllm --model <model> --disaggregation-mode prefill
$
$
# Decode worker (different NIXL port to avoid collision)
$
NIXL_TELEMETRY_ENABLE=y NIXL_TELEMETRY_EXPORTER=prometheus \
>
  NIXL_TELEMETRY_PROMETHEUS_PORT=19091 DYN_SYSTEM_PORT=8082 \
>
  python -m dynamo.vllm --model <model> --disaggregation-mode decode
$
$
# Scrape NIXL metrics (separate from Dynamo metrics on 8081/8082)
$
curl http://localhost:19090/metrics

For the full list of metrics, configuration options, and architecture details, see the upstream NIXL Telemetry documentation and Prometheus exporter README. For Kubernetes, see Enable NIXL Telemetry.

Related Documentation

• Distributed Runtime Architecture
• Dynamo Architecture Overview
• Backend Guide