Automated maintenance windows via custom tagging

[robhalli]

Hi everyone,

We’ve implemented a continuous maintenance window that targets any host with the custom tag Maintenance:true.

This approach allows us to place hosts into maintenance dynamically and automatically using the custom tagging API. It eliminates the need to repeatedly create or modify maintenance windows via the API and instead gives us a simple on/off toggle for individual hosts or groups of servers.

This works perfectly for host‑based metric and process alerts (CPU, memory, storage, process availability, etc.), as those alerts are correctly suppressed.

However, we are still receiving Davis anomaly events and synthetic alerts from hosts that have the Maintenance:true tag applied.

My question:
Is it possible to have Davis anomaly events and synthetic alerts automatically included in the maintenance window without explicitly defining them? Logically, if the affected host is within a maintenance window, we would expect all events and alerts generated by that host—regardless of type—to be suppressed.

Example scenario:

• Host: host1.example.com
• Custom tag: Maintenance:true
• Result: No host-based metric or process alerts (expected), but Davis anomaly events and synthetic alerts are still being generated.

Please refer to the attached diagram for a high‑level overview of our current process.

Thanks in advance for any guidance or best practices.

[AntonPineiro]

Hi,

Host, processes and services have a topology relation to propagate tag between them.

I do not think Synthetic relation with that entities is done automatically, you need to "link" those Synthetics to entities or create some rule about it.

Maybe workflow is a good candidate, something as if entity X or Y have tag "Maintenance:true", include that tag also to Synthetic A or B.

And workflow should remove Synthetic tag when it is removed en entities X or Y.

Best regards