Master Cotonity's activity log interface to trace execution paths, inspect step inputs and outputs, filter by run status and time, and export log data for offline analysis.
Navigating the activity log
The Activity tab on any agent page is your primary debugging tool. It lists all execution runs in reverse chronological order, with columns showing the run ID, trigger type, start time, duration, and final status. Use the filter bar to narrow the list by status (All, Completed, Failed, Timed Out), trigger type (Scheduled, Webhook, Manual), and time range. When investigating a specific incident, filter by the time window during which the problem occurred to avoid searching through hundreds of unrelated runs. The run ID shown in the list is stable and can be shared with the support team if you need to escalate an issue.
Inspecting step inputs and outputs
Clicking any run opens the execution trace, which shows every step in the order it executed. Click a step to open its detail panel. The Input tab shows the data the step received from upstream steps and the agent's context — this is what the step worked with. The Output tab shows the data the step produced — this is what downstream steps received. Comparing the input and output for a failing step often immediately reveals the problem: a missing field, an unexpected null, or a type mismatch. For HTTP request steps, the detail panel also shows the full request URL, headers, and body that were sent, and the raw HTTP response including status code, headers, and body.
Adding log steps for custom instrumentation
When the standard activity log does not provide enough detail to diagnose a subtle bug, add a Log step at strategic points in the workflow. A Log step writes a custom message and any variable values you specify to the run's activity log without performing any external action. This is useful for tracking the state of complex computed values, recording which branch a conditional took, or confirming that a data transformation produced the expected result. Log steps are visible in the execution trace alongside regular steps. To avoid unnecessary verbosity in production, wrap Log steps in a conditional that checks a 'debug mode' flag, and disable the flag once the issue is resolved.
Exporting and searching log data
For workflows with high run volumes, the in-app activity log's 30-day retention window and page-based navigation may be insufficient for thorough debugging. Cotonity supports exporting activity log data in JSON or CSV format from the Activity tab's export menu. You can filter the export by date range, status, and agent. The exported data includes all step inputs and outputs, making it suitable for analysis in a spreadsheet or a SQL query tool. Enterprise accounts can also configure a log streaming integration under Settings > Integrations > Log Streaming to pipe run data in real time to an external observability platform such as Datadog or Splunk.