Orkes Conductor MCP. Track Every Microservice Workflow State.

Look, you gotta know what's running in your microservice pipeline without opening some heavy dashboard and getting lost. This MCP Server connects your AI agent right into the Orkes Conductor workflow orchestration layer. It gives you full visibility—you can check definitions, watch things run live, and deep-dive into exactly how any process worked.

Defining the Workflows:
You don't wanna guess what processes exist. You can start by getting a complete list of every overarching workflow definition available in the Conductor cluster using list_workflow_defs. This shows you all the blueprints, giving you a high-level overview of what your system is capable of handling. If you need to know exactly how one of those specific blueprints works—like its exact schema or graph structure—you can run get_workflow_def against any named definition.

That pulls up the full map, showing every task and branching point built into that workflow.

Checking Component Availability:
Before you build anything, you gotta know what tools are available to build with. To see every individual operation—the callable tasks registered in your system—you use list_task_defs. This gives you the complete catalog of actions; it tells you exactly what kind of operations can be plugged into any workflow definition.

Monitoring Live Runs:
If something's running right now, you wanna know about it. You don't have to manually check status pages. Using list_running, your agent lists every active instance of a specific workflow that’s currently executing and progressing through its steps. This lets you monitor live activity against the name of the blueprint.

Digging Into History:
When something breaks, or when you just need to prove something worked last week, you gotta check the history. Instead of sifting through logs manually, you can use search_workflows. This function performs a flexible search across thousands of past executions. You can filter those results using status—did it pass or did it fail?—a specific ID number, or even by workflow type to find exactly what you're looking for in the Elasticsearch index.

For an even deeper look at just one single run, get_execution is your ticket. This retrieves everything about a single workflow instance: its full final state, the inputs and outputs that were passed between every task step, and the detailed task trace history showing exactly what happened when it ran.

You don't just get a status; you get the whole receipt.

Basically, this setup lets your agent interact with all the core functions of Orkes Conductor—you can list definitions with list_workflow_defs, pull up a specific blueprint's graph using get_workflow_def, see every single task available via list_task_defs, check which workflows are running right now with list_running, search across all history records using search_workflows, and finally, get the full operational truth for any one run with get_execution.