Discuss with expertTry it for free
© 2026 Direct Cursus Technology L.L.C.

Yandex Cloud AI Studio metrics

Written by
Updated at September 17, 2026
View in Markdown

This section describes the AI Studio metrics delivered to Monitoring.

The name label is for the metric name.

MCP Hub metrics

Labels shared by all metrics of the AI Studio MCP Hub feature:

Label Value
service Service ID: serverless-mcp-gateway
mcp_gateway_id MCP server ID

Service metrics:

Name
Type, units

Description
Labels

requests_completed_per_second
DGAUGE, requests per second

Number of completed requests to the MCP server per second.
Labels:

  • request_text: Type of your request.
  • error_code: Error, if any.

requests_received_per_second
DGAUGE, requests per second

Number of requests to the MCP server per second.
Label:

  • request_text: Type of your request.

sessions_active_count
DGAUGE, count

Number of open sessions with the MCP server.

sessions_closed_per_second
DGAUGE, requests per second

Number of session closures with the MCP server per second.

sessions_duration_milliseconds
IGAUGE, milliseconds

Session duration with the MCP server.
Label:

  • bin: Histogram buckets.

sessions_init_requests_per_second
DGAUGE, requests per second

Number of session initialization requests to the MCP server per second.
Label:

  • error_code: Error, if any.

tool_call_duration_milliseconds
IGAUGE, milliseconds

Execution duration of a request to the MCP server's tool.
Label:

  • bin: Histogram buckets.

Yandex Workflows metrics

All Workflows metrics share the workflow label, which is the workflow ID.

Metric name
Type, units		 Description
Labels
step_completed_per_second
DGAUGE, steps per second		 Number of completed steps per second.
This metric features the step label containing the step name.
step_failed_per_second
DGAUGE, steps per second		 Number of failed steps per second.
Labels:
  • step: Step name
  • error_code: Error code.
step_inflight_count
DGAUGE, steps		 Number of steps currently in progress.
This metric features the step label containing the step name.
step_started_per_second
DGAUGE, steps/s		 Number of started steps per second.
This metric features the step label containing the step name.
step_time_milliseconds
DGAUGE, milliseconds		 Step execution time.
Labels:
  • bin: Histogram buckets.
  • step: Step name
step_time_milliseconds_count
DGAUGE, number of steps		 Total number of step time measurements.
This metric features the step label containing the step name.
step_time_milliseconds_sum
DGAUGE, milliseconds		 Total execution time for all steps.
This metric features the step label containing the step name.
workflow_completed_per_second
DGAUGE, workflows per second		 Number of completed workflows per second.
workflow_failed_per_second
DGAUGE, workflows per second		 Number of failed workflows per second.
This metric features the error_code label containing the error code.
workflow_inflight_count
DGAUGE, workflows		 Number of workflows currently in progress.
workflow_scheduled_per_second
DGAUGE, workflows per second		 Number of scheduled workflows per second.
workflow_started_per_second
DGAUGE, workflows per second		 Number of started workflows per second.
workflow_time_milliseconds
DGAUGE, milliseconds		 Execution time per workflow.
This metric features the bin label containing histogram buckets.
workflow_time_milliseconds_count
DGAUGE, count		 Total number of workflow time measurements.
workflow_time_milliseconds_sum
DGAUGE, milliseconds		 Total execution time for all workflows.
Previous
All metric references
Next
API Gateway