# tensorplay.distributed.elastic.metrics API

Source: https://www.tensorplay.cn/docs/api/tensorplay.distributed.elastic.metrics.html

## Functions 2

[#](#api-tensorplay.distributed.elastic.metrics.configure)

### configure

function[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.metrics.configure.html)

```python
tensorplay.distributed.elastic.metrics.configure(handler: MetricHandler, group: str | None = None) → None
```

Route group (or all groups) to handler.

[#](#api-tensorplay.distributed.elastic.metrics.prof)

### prof

function[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.metrics.prof.html)

```python
tensorplay.distributed.elastic.metrics.prof(fn=None, group: str = 'tp_elastic', interval_ms: float = 1000)
```

Decorator measuring wall time of fn and emitting it as a metric.

Works both bare (@prof) and parameterized (@prof(group=...)).

## Classes 1

[#](#api-tensorplay.distributed.elastic.metrics.MetricsConfig)

### MetricsConfig

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.metrics.MetricsConfig.html)

```python
class tensorplay.distributed.elastic.metrics.MetricsConfig(params: dict[str, str] | None = None, default_handler: str = 'console', cfg: dict[str, str] | None = None)
```

Metric plumbing configuration; maps groups to handler names.