Avoiding race conditions for advanced agents

To avoid race conditions associated with custom actions for advanced agents, make the actions thread-safe by ensuring that the agent runs on one node at a time.

Agent rules are no longer being actively developed; Queue Processor rules replace standard agents. For more information, see Use Job Scheduler and Queue Processor rules instead of agents.

  1. In the navigation pane of Admin Studio, click Resources > Agents.
  2. Click the name of an agent that shows exceptions to open the agent in Dev Studio.
  3. For each agent, expand the agent record to display the Advanced area.
    Note: If you see a warning that agents will be deprecated, you can still modify the agent as described here.
  4. Configure the Advanced settings:
    • Select the Run this agent on only one node at a time check box.
    • Configure the agents to run on one node at a time at staggered intervals by selecting the Delay next run of agent across the cluster by specified time period check box.
    • Optional: Enter a time window during which to block agent actions in the Block execution within this time window fields.
  5. Click Save.
  6. Repeat steps 2 through 5 for each agent which was causing exceptions.
  7. Retest your application in a multi-node system.