Observability & Debugging

Execution Logs

Every workflow execution produces a complete, immutable audit trail — allowing you to debug failures, analyze performance, and understand agent behavior at every step.

Why Execution Logs Matter

In AI automation, failures are inevitable — models hallucinate, APIs timeout, tools misbehave. This platform treats observability as a first-class feature. Logs are not an afterthought; they are the foundation for trust, safety, and reproducibility.

Observability Stack

Run History

Every workflow trigger creates a permanent execution record.

Step-Level Status

Track success, failure, retries, and timing for each step.

Payload Inspection

Full visibility into inputs, outputs, and intermediate results.

Log Structure

Logs are stored per task execution and indexed by workflow, step, and timestamp. This enables efficient querying, debugging, and replay.

FieldDescriptionExample
trace_idUnique identifier for a full workflow executiontr_8273ab91
step_idThe workflow step being executedemail_send_01
statusExecution result of the stepSUCCESS | FAILED | RETRY
latencyExecution time in milliseconds482ms
payloadInput parameters and output results{ input, output }

Where Logs Are Stored

Execution logs are persisted locally in MongoDB under each task record. They are indexed by workflow ID, task ID, step ID, and timestamp.

Failure Debugging Workflow

When a workflow fails, inspect the failed step’s log entry to view the exact error, model output, and tool response. You can replay executions with identical context to validate fixes or tune prompts without re-triggering upstream steps.

Trust & Compliance

Because all logs are stored locally and never leave your environment, this system provides stronger privacy guarantees than cloud-hosted automation platforms. This is critical for sensitive workflows and regulated data.