criticalRecovery and warningRecovery always fail for createMonitor

[paco-sparta]

Describe the bug
This is a problem with the backend, but I cannot find the right repo.

I have a project with the following monitor, exported directly from the UI:

  type: MonitorType.QUERY_ALERT,
  query: `avg(last_1d):anomalies(max:system.cpu.user{env:${enviro},service:${monitor.service}} by {host}, 'agile', 2, direction='both', interval=300, alert_window='last_1h', count_default_zero='true', seasonality='weekly') > 0.5`,
  message: `[${monitor.service}] CPU usage is too high.`,
  options: {
    thresholds: {
      critical: 0.5,
      criticalRecovery: 0.25,
      warning: 0.2,
      warningRecovery: 0.1,
    },

For this use case the API monitorsApi.createMonitor always returns:

"error": {
      "code": 400,
      "body": {
        "errors": [
          "alert and alert recovery criteria are such that the monitor can be simultaneously in alert and alert recovery states"
        ]
      }
    }

No matter what values I put in criticalRecovery and warningRecovery both always error.

To Reproduce
Steps to reproduce the behavior:

1. Using the API
2. Create a QUERY_ALERT monitor with criticalRecovery and warningRecovery
3. Request createMonitor
4. See error

Expected behavior
These values should be accepted, as they match the website's.

Environment and Versions (please complete the following information):
"@datadog/datadog-api-client": "^1.0.0",

Additional context
This is a backend problem, the error happens after the request goes through. I'd be happy to forward it to the corresponding repo.

[therve]

Hi,

Can you submit a ticket with support please: https://www.datadoghq.com/support/. There is no GIthub repo for backend. Thanks.

[paco-sparta]

Is there a way not to go through support escalations for what is more than probably just a code bug you can ask about in an internal channel?

[paco-sparta]

The solution was to change triggerWindow and recoveryWindow on the options.thresholdWindows object so that trigger <= recovery

    thresholds: {
      critical: 0.5,
      criticalRecovery: 0.25,
      warning: 0.2,
      warningRecovery: 0.1,
    },
    thresholdWindows: {
      triggerWindow: 'last_30m',
      recoveryWindow: 'last_30m',
    },