> This page is for NeMo Fabric, version Latest (v0.6.0) (default).
> For other versions, use one of these documentation indexes:
> - Latest (v0.6.0) (default): https://docs.nvidia.com/nemo/fabric/latest/llms.txt
> - dev: https://docs.nvidia.com/nemo/fabric/dev/llms.txt
> - v0.6.0: https://docs.nvidia.com/nemo/fabric/v0.6.0/llms.txt
> - v0.5.0: https://docs.nvidia.com/nemo/fabric/v0.5.0/llms.txt
> - v0.4.0: https://docs.nvidia.com/nemo/fabric/v0.4.0/llms.txt
> - v0.3.0: https://docs.nvidia.com/nemo/fabric/v0.3.0/llms.txt
> - v0.2.0: https://docs.nvidia.com/nemo/fabric/v0.2.0/llms.txt
> - v0.1.1: https://docs.nvidia.com/nemo/fabric/v0.1.1/llms.txt
> - v0.1.0: https://docs.nvidia.com/nemo/fabric/v0.1.0/llms.txt

> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/fabric/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/fabric/_mcp/server.

# Struct Invocation System Stats

> System measurements around one invocation, collected inside the executing runner. Process-tree observations cover its adapter host and descendants, excluding the orchestrator and remote model/service processes. The separate cgroup observations may include other workloads in the same cgroup.

Generated from `cargo doc --no-deps -p nemo-fabric-core`.

<pre />

System measurements around one invocation, collected inside the executing runner. Process-tree observations cover its adapter host and descendants, excluding the orchestrator and remote model/service processes. The separate cgroup observations may include other workloads in the same cgroup.

Endpoints are instantaneous readings, not peaks or time averages. CPU time is cumulative since runtime startup; its difference is only an estimate because process reaping and reparenting can change which work is visible.

## Fields

### `collected_groups: Vec<SystemStatsMetricGroup>`

Requested groups, including selected groups unavailable on this platform.

### `elapsed_seconds: f64`

Monotonic elapsed seconds for the adapter exchange. Excludes runtime startup, endpoint collection, artifact preparation, and teardown. Optional sampling runs concurrently; its effect is not subtracted.

### `before: ResourceObservation`

Observation immediately before the adapter exchange.

### `after: ResourceObservation`

Observation immediately after the adapter exchange, before another invocation can acquire this runtime's host lock.

### `cpu_time_seconds_estimate: Option<f64>`

Nonnegative CPU counter difference, in seconds. Absent if an endpoint is missing, the runtime identity differs, or its CPU counter decreased. It can undercount work whose processes leave the tree between reads.

### `cgroup: Option<CgroupStats>`

Before/after readings for the adapter host's cgroup. This scope may contain other workloads and overlaps the process tree; never add them.

### `sampled: Option<SampleSummary>`

Bounded sample reductions and collection quality when sampling is enabled.

## Trait Implementations

### `impl Clone for InvocationSystemStats`

<pre />

#### `clone`

<pre />

#### `clone_from`

<pre />

### `impl Debug for InvocationSystemStats`

<pre />

#### `fmt`

<pre />

### `impl<'de> Deserialize<'de> for InvocationSystemStats`

<pre />

#### `deserialize`

<pre />

### `impl JsonSchema for InvocationSystemStats`

<pre />

#### `schema_name`

<pre />

#### `schema_id`

<pre />

#### `json_schema`

<pre />

#### `inline_schema`

<pre />

### `impl PartialEq for InvocationSystemStats`

<pre />

#### `eq`

<pre />

#### `ne`

<pre />

### `impl Serialize for InvocationSystemStats`

<pre />

#### `serialize`

<pre />

### `impl StructuralPartialEq for InvocationSystemStats`

<pre />