TestMesh
Dashboard

Executions

Read execution results, monitor live step progress, inspect response details, and diagnose failures with AI-powered root cause analysis.

Every time you run a flow — manually, on a schedule, or via the CLI — an execution record is created. The Executions page gives you real-time and historical visibility into what happened.

Execution list

The Executions page shows all recent runs with:

  • Flow name and workspace
  • Triggered by (user, schedule, API, CLI)
  • Start time and duration
  • Overall status: passed, failed, running, cancelled

Click any execution to open the detail view.

Execution detail

The execution detail view has two panels:

Left — Step timeline: A vertical list of all steps in the order they executed. Each step shows:

  • Step ID and action type
  • Status icon: ✓ passed, ✗ failed, ⚡ skipped, ↻ retrying
  • Duration

Click a step to open its details in the right panel.

Right — Step details: Shows the full input and output for the selected step, formatted by action type:

Action typeWhat you see
http_requestMethod, URL, request headers + body, response status, response headers + body
database_queryConnection (masked), SQL query, parameters, returned rows
kafka_producerTopic, key, value, partition and offset
kafka_consumerMessages received, each with key, value, offset, timestamp
redis_get / redis_setKey, value, TTL
grpc_callService, method, request, response

Assertion results are shown below the output — each assertion expression with its evaluated value and pass/fail.

Output variable extractions are shown last — the JSONPath expressions and the values they captured.

Live execution monitoring

When a flow is actively running, the execution detail view streams updates over WebSocket. Steps appear in real time as they complete. The page does not need to be refreshed.

If you navigate away during a run, the execution continues in the background. Return to the Executions page to find it.

Execution status values

StatusMeaning
runningFlow is actively executing
passedAll steps completed and all assertions passed
failedOne or more steps failed or assertions did not pass
cancelledManually stopped before completion
timed_outFlow-level timeout was exceeded

A flow with skipped steps (via run_condition) is still passed if all executed steps passed.

Failure inspection

When a step fails, the step detail panel shows exactly what went wrong:

Assertion failures — each failed assertion is highlighted with:

  • The expression that was evaluated
  • The actual value that was present
  • What was expected

Action errors — for connection errors, timeouts, or action-level failures, the raw error message is shown with the partial response if one was received.

Retry history — if a step was configured with retries, each attempt is shown with its own result. You can see which attempt succeeded or whether all attempts failed.

AI failure analysis

For failed executions, click Analyse with AI in the execution header. The AI examines:

  • Which step failed and why
  • What the actual vs expected values were
  • Patterns from prior executions of the same flow
  • Potential root causes: service errors, environment issues, data problems, timing issues

The analysis appears inline. It includes a suggested next action — whether that's checking a service log, adjusting a timeout, or updating an assertion.

Re-running an execution

From any execution detail view, click Re-run to execute the same flow again with the same environment and variable overrides. Useful for confirming a fix without navigating back to the Flows page.

The new execution is linked to the original — you can switch between them from the execution header.

Execution history and retention

All execution records are stored in PostgreSQL. The Executions list shows the 100 most recent by default. Use the date filter to search further back.

Execution records include the full step input/output at the time of the run. If you edit a flow after a run, the execution record reflects what the flow looked like when it ran, not the current version.

Analytics

Aggregate execution data is available in the Analytics section:

  • Pass rate trends over time
  • Flakiness detection (steps that pass and fail intermittently)
  • Duration trends per flow
  • Endpoint coverage (which services and routes are exercised by your flows)
  • AI-powered diagnosis of recurring failures

On this page