Appearance
Core Concepts
Automation Watchdog is easiest to understand once you separate the product into a few core ideas.
The core building blocks
| Concept | What it means |
|---|---|
| Watch | A monitoring definition for a workflow or operational expectation |
| Check-in | The API request a workflow sends to report progress |
| Condition | The monitored state for the watch, queue, or machine being tracked |
| Activation | The event that starts a watch or condition monitoring |
| Error reason | A specific reason the condition is currently unhealthy |
Watch types
Primary mode
Missed Action watch
Expects check-ins and enters error when expected activity is missing or quality rules are broken.
Simple mode
Notify Only watch
Relays a notification when a check-in arrives, without monitoring for missed check-ins.
Check-ins are the signal
A check-in is the API call your workflow makes to Automation Watchdog. A check-in can be simple, or it can carry context such as:
- queue
- machine
- outcome
That lets a watch monitor more than simple uptime. It can monitor the right queue, the right worker, or the quality of the recent run.
Conditions and aggregation
Depending on the monitoring mode, a watch may track:
- one shared condition for the whole watch
- one condition per queue
- one condition per machine
Those individual conditions roll up into the overall watch state shown in the application.
Environments matter
Teams often use different API tokens and environment flags for development, test, and production. This helps keep monitoring behavior aligned with where the workflow is running and where alerts should be sent.