devops-infra MCP Server
MCP server for Celery Flower — monitor workers, manage tasks and queues from any AI assistant
Discovered via github-topic:mcp and last synced 3mo ago.
1. Install the package
uvx celery-flower-mcp
2. Add to claude_desktop_config.json
{
"mcpServers": {
"celery-flower-mcp": {
"command": "npx",
"args": [
"celery-flower-mcp"
]
}
}
}Config file location: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) / %APPDATA%\Claude\claude_desktop_config.json (Windows)
Default
Gracefully shut down a worker
List all registered task types across workers
Set rate limit for a task on a worker (e.g. `100/m`)
Restart a worker's process pool
Get full details for a task by UUID
`http://localhost:5555`
Remove N processes from a worker's pool
Execute a task synchronously and wait for the result
Check whether the Flower instance is reachable and healthy
—
Make a worker start consuming from a queue
Send a task by name — no registration required on worker side
—
Make a worker stop consuming from a queue
Abort a running task
List all workers — optionally filter by name, refresh live stats, or get status only
List tasks with filters: state, worker, name, date range, search, pagination
Set soft and/or hard time limits for a task on a worker
—
Configure autoscale min/max bounds
Dispatch a task asynchronously, returns task UUID
Description
Revoke a task; optionally terminate with a signal
Add N processes to a worker's pool
Retrieve a task's result (with optional timeout)
Get the current depth of all configured queues