nat.runtime.metrics

Per-user workflow resource usage monitoring models and collector.

Attributes

logger

Classes

PerUserSessionMetrics	Session lifecycle metrics for a per-user workflow.
PerUserRequestMetrics	Request-level metrics for a per-user workflow.
PerUserMemoryMetrics	Memory/resource count metrics for a per-user workflow.
PerUserResourceUsage	Combined resource usage metrics for a single per-user workflow.
PerUserMonitorResponse	Response model for the /monitor/users endpoint.
PerUserMetricsCollector	Collector for per-user workflow metrics.

Module Contents

logger

class PerUserSessionMetrics(/, **data: Any)

Bases: pydantic.BaseModel

Session lifecycle metrics for a per-user workflow.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

created_at: datetime.datetime = None

last_activity: datetime.datetime = None

ref_count: int = None

is_active: bool = None

class PerUserRequestMetrics(/, **data: Any)

Bases: pydantic.BaseModel

Request-level metrics for a per-user workflow.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

total_requests: int = None

active_requests: int = None

avg_latency_ms: float = None

error_count: int = None

class PerUserMemoryMetrics(/, **data: Any)

Bases: pydantic.BaseModel

Memory/resource count metrics for a per-user workflow.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

per_user_functions_count: int = None

per_user_function_groups_count: int = None

exit_stack_size: int = None

class PerUserResourceUsage(/, **data: Any)

Bases: pydantic.BaseModel

Combined resource usage metrics for a single per-user workflow.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

user_id: str = None

session: PerUserSessionMetrics = None

requests: PerUserRequestMetrics = None

memory: PerUserMemoryMetrics = None

class PerUserMonitorResponse(/, **data: Any)

Bases: pydantic.BaseModel

Response model for the /monitor/users endpoint.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

timestamp: datetime.datetime = None

total_active_users: int = None

users: list[PerUserResourceUsage] = None

class PerUserMetricsCollector(

session_manager: nat.runtime.session.SessionManager,

)

Collector for per-user workflow metrics.

This class aggregates metrics from SessionManager’s per-user builders and provides methods to collect metrics for individual users or all users.

Initialize the collector with a SessionManager reference.

Args:

session_manager: The SessionManager instance to collect metrics from

_session_manager

async collect_user_metrics(user_id: str) → PerUserResourceUsage | None

Collect metrics for a specific user.

Args:

user_id: The user identifier to collect metrics for

Returns:

PerUserResourceUsage if user exists, None otherwise

async collect_all_metrics() → PerUserMonitorResponse

Collect metrics for all active per-user workflows.

Returns:

PerUserMonitorResponse with all user metrics

_build_user_metrics(

user_id: str,

builder_info,

) → PerUserResourceUsage

Build metrics for a single user from builder info.

Args:

user_id: The user identifier builder_info: The PerUserBuilderInfo instance

Returns:

PerUserResourceUsage with all metrics