Retrieves historical Data Integration activity logs, session logs, and error logs from the Informatica IDMCMonitor service.
The Log Management server exposes IDMC's read-only historical logging endpoints to AI agents. Use this server to enables agents to browse completed Data Integration jobs and retrieve session and error logs. This MCP server provides historical data. It never starts, stops, or changes the state of a job or taskflow, and it doesn't poll live activity
Key capabilities
This MCP server offers the following capabilities:
•Completed-job search. Searches completed Data Integration activity log entries, optionally narrowed to a specific task, run, or log entry ID.
•Error log retrieval. Retrieves the error log for a specific run.
•Session log retrieval. Retrieves the session log for a specific activity log entry. Includes logs from the child tasks of a multi-task job.
Use cases
Use this MCP server to address the following use cases:
•Failed job root cause. Diagnose failed runs by locating their failed-state entries and error messages in the activity log, then review the associated session log and error log for row-level context and rejected records.
•Run history retrospective. Summarize a task's completed jobs by searching the activity log and returning the state outcomes, source and target row counts, and transformation throughput.
•Clean run sign-off. Confirm that a completed job finished successfully by reviewing the associated activity log and error log.
•Rejected records follow-up. Read error logs for jobs that completed with errors and compile a correction list that describes why records in rejected rows didn't load.
Tools
The following table describes the tools available on this MCP server:
Tools
Description
search_activity_log
Searches activity log entries for completed Data Integration jobs. All parameters are optional. You can filter by log entry ID or task. When used together, taskID and runID search for a specific completed job.
get_session_log
Retrieves the session log for a specific activity log entry. If you specify an item ID or child item ID, retrieves the session log for that item. For a single-task job, the response is in plain text. For a multi-task job, supplying the item ID and child item ID targets the session log for a specific subtask.
get_error_log
Retrieves the error log for a specific activity log entry. An empty response body indicates that the job completed without any error records.
Note:
In AI Agent Engineering, the MCP server tools are called "actions."