Workflows

View, manage, and monitor all your registered AI workflows.

What are Workflows?

A workflow is a logical grouping of traces that represent the same AI application or agent. When you instrument your code with @instrument(name="..."), TuringPulse automatically registers that workflow.

Workflows Catalog

Navigate to Operations → Workflows to see all registered workflows:

  • Workflow Name - The agent_id from your instrumentation
  • Status - Active, Inactive, or Error state
  • Run Count - Total executions
  • Last Run - When it was last executed
  • Health - KPI, drift, and anomaly indicators

Workflow Detail Page

Click on any workflow to see detailed information:

Overview Tab

  • Key metrics (latency, throughput, error rate)
  • Recent runs list
  • Active alerts and incidents

KPIs Tab

  • KPI rules configured for this workflow
  • Current KPI values and trends
  • Alert history

Drift Tab

  • Drift detection rules
  • Baseline vs current distributions
  • Drift events history

Anomalies Tab

  • Anomaly detection rules
  • Detected anomalies list
  • Anomaly trends

Settings Tab

  • Workflow configuration
  • Labels and metadata
  • Notification preferences

Workflow Actions

From the workflow detail page, you can:

  • View Runs - See all executions for this workflow
  • Configure KPIs - Set up threshold alerts
  • Configure Drift - Set up baseline monitoring
  • Configure Anomalies - Set up outlier detection
  • Edit Settings - Update workflow metadata

Filtering Workflows

Use the search and filter options to find workflows:

  • Search - Filter by name
  • Status - Active, Inactive, Error
  • Labels - Filter by custom labels
  • Health - Show only workflows with issues

Next Steps