datadog_monitor ignore query alert vs metric alert diff

[nmuesch]

Fixes #241

The API will return query alert or metric alert depending on the query. We currently ignore the fact that the user may have a metric alert but the API returns a query alert but we don't ignore the opposite.

This config seems to replicate the behavior:

  name = "Anomolous CPU usage"
  type = "query alert"
  message = "CPU utilization is outside normal bounds"
  query = "avg(last_4h):anomalies(ewma_20(avg:system.cpu.system{env:prod,service:website}.as_rate()), 'robust', 3, direction='below', alert_window='last_30m', interval=60, count_default_zero='true', seasonality='weekly') >= 1"
  thresholds {
    critical          = 1.0
    critical_recovery = 0.0
  }
  threshold_windows {
    trigger_window    = "last_30m"
    recovery_window   = "last_30m"
  }

  notify_no_data    = false
  renotify_interval = 60
}

This monitor is seen by Datadog as a metric alert. So when the read function occurs the provider sees this as a change in type and does a forceNew. This PR aims to suppress that diff.

[bkabrda]

LGTM, thanks for doing this PR!