Monitoring Modes

Automation Watchdog lets you choose how monitoring conditions should be grouped. The right mode depends on what counts as "healthy" for your workflow.

Standard mode

Use standard mode when any valid check-in can satisfy the watch timing.

This is the simplest mode. A shared condition tracks the whole watch. If one machine checks in, that shared condition is satisfied.

Condition By Queue

Use Condition By Queue when the same workflow can operate against multiple queues and each queue should be monitored independently.

• each queue gets its own condition
• each queue can become healthy or unhealthy independently
• the overall watch aggregates those queue conditions

Condition By Machine

Use Condition By Machine when each machine should be monitored independently.

• each machine gets its own condition
• each machine has its own timing and error state
• the overall watch is healthy only when all active machine conditions are healthy

Expected Machine Count

Expected Machine Count adds another rule for workflows that should fan out across a known number of workers. It is useful when the watch should not clear its overdue state until enough machines have reported in the current run.

Which mode should you choose?

If your goal is...	Best fit
Any machine checking in is enough	Standard mode
Each queue should be monitored separately	Condition By Queue
Each machine must independently meet check-in timing	Condition By Machine
A run should involve a minimum number of workers	Standard or Condition By Machine with Expected Machine Count

Pair mode with timing

Monitoring mode answers what gets its own condition. Timing strategy answers how each condition decides whether it is overdue.

That means you choose both:

• Standard, queue, or machine mode
• Fixed or cascading period

Read next

• Outcomes and Thresholds
• Multi-Machine Workers scenario