Skip to main content

Manage agents and agent pools

In briefAgents run your jobs, and agent pools route work to whichever agent is available. The Agents page holds every agent, pool and relay, and lets you mark an agent out of service for maintenance. A Universal Agent pool is a queue its agents claim work from; a legacy agent group is a set of legacy agents of one type that a legacy job can target.

Notes
  • A pool job goes on the pool's queue once, and the first agent to ask for it runs it. A job for one specific Universal Agent can be taken by any agent in that agent's pool.
  • A Universal Agent obeys its own MAX_CONCURRENT_JOBS setting (50 unless changed), not the cap shown on its row.
  • Unknown on a legacy agent means its relay can't vouch for it, not that the agent is down.
  • A mark survives a reconnect and holds jobs that target the agent; Clear Mark releases them.
  • Mark Offline stops communication with the agent, and Clear Mark opens a new connection; Mark Draining keeps it connected.
  • Register a Universal Agent through the API; the install commands New Agent shows for Universal don't register it. A new one reads Offline until its program first checks in.
  • Machine Properties on an agent's Edit Agent dialog holds that machine's own values, read by [[MI.Name]]. They are not encrypted.
Watch out
  • A legacy agent's Default event environment left unset means every event that machine raises is lost.
  • A legacy agent with no FT Role takes part in no file transfer.
  • A pool can't be deleted while it has agents.
  • Agents and Agent Pools permissions count only under All workspaces; a grant confined to particular workspaces hides the tab.
  • Allow Kill is off by default on an SAP R/3 and CRM agent, so its jobs cannot be killed until you turn it on.
  • Use TLS for Scheduling never falls back to an unencrypted connection: an agent whose own TLS is off stays offline.