Skip to main content
Every monitor has a frequency that determines how often the hub schedules it for execution.

​
Setting a frequency

Frequency is a required field in both builders: 
import { Frequency } from "@griffin-app/griffin";

createMonitorBuilder({
  name: "health-check",
  frequency: Frequency.every(5).minutes(),
})

​
Available intervals

Frequency.every(1).minute()    // Every minute
Frequency.every(5).minutes()   // Every 5 minutes
Frequency.every(15).minutes()  // Every 15 minutes

Frequency.every(1).hour()      // Every hour
Frequency.every(6).hours()     // Every 6 hours

Frequency.every(1).day()       // Every day
Frequency.every(7).days()      // Every 7 days
Both singular and plural forms work: .minute() and .minutes() are equivalent.

​
How scheduling works

The hub’s scheduler runs on a tick interval (default: 30 seconds). On each tick, it checks which monitors are due based on their frequency and last execution time, then enqueues jobs for execution.
  • A monitor with Frequency.every(1).minute() will be scheduled approximately every 60 seconds
  • The scheduler ensures no duplicate jobs are created for the same monitor within its frequency window
  • If an executor is unavailable, jobs queue up and execute when an executor becomes available

​
Choosing a frequency

FrequencyGood for
Every 1 minuteCritical health checks, payment APIs, auth endpoints
Every 5 minutesImportant API endpoints, core business flows
Every 15 minutesStandard monitoring, CRUD operations
Every 1 hourLow-priority checks, third-party API status
Every 1 dayPeriodic validation, data integrity checks
Higher frequencies mean more executions, which affects resource usage on self-hosted deployments and costs on Griffin Cloud.

​
Locations

Optionally specify where your monitor should execute. In distributed mode, jobs are routed to executors running in the specified locations: 
createMonitorBuilder({
  name: "global-check",
  frequency: Frequency.every(5).minutes(),
  locations: ["us-east-1", "eu-west-1"],
})
When locations are specified, the scheduler creates separate jobs for each location on each run. If no locations are specified, the monitor runs on any available executor.