Fix the following issue:
Support scheduling_options / cumulative evaluation window on DatadogMonitor
I'm using the operator to manage monitors via GitOps. I'd like to set up a few "calendar-aligned" cumulative monitors — e.g. a monthly cost / usage budget alert that resets on the 1st of each month, and a daily SLO-like accumulator that resets at 00:00 UTC.
In the Datadog UI (and via the public Datadog API) this is exposed through scheduling_options.evaluation_window, where you can pick a cumulative window aligned to the start of the hour, day, or month (hour_starts / day_starts / month_starts). The behaviour is different from a rolling timeframe — values accumulate from the alignment point instead of sliding.
The problem is I don't see any way to express this on a DatadogMonitor CR. spec.options covers a lot of the usual stuff (timeoutH, requireFullWindow, notificationPresetName, thresholds, renotify settings, etc.) but there's no field for the evaluation window / scheduling options. If I add it under spec.options anyway, the K8s API server rejects it as an unknown field, and even if I bypass that the controller has no logic to forward it to the Datadog API — the resulting monitor in Datadog ends up with the default rolling window instead of the cumulative one I asked for.
For folks managing budget / quota style monitors as code, this means we currently have to click those monitors together in the UI (or maintain them through a separate Terraform pipeline) instead of keeping them next to the rest of our DatadogMonitor manifests.
Could the operator expose the scheduling options / cumulative evaluation window on DatadogMonitor so it reaches parity with the Datadog API? All three alignment kinds (hourly / daily / monthly) should be configurable, and the controller should pass them through when creating / updating the monitor.