← sources

Do you automatically isolate servers/devices based on detetctions?

source: reddit (reddit:sysadmin)published 2026-09-05 18:28ingested 2026-09-06 02:40

retained text

We don't have a 24/7 SOC, so we are thinking about automatically isolating servers and some high-value devices based on custom Defender for Endpoint detections.  Obviously, we want to do that only for high-precision and high-confidence detections, such as opening a shell from a strange parent process.

If we got one of these detections during working hours, there would be someone to react.  But after about 7 PM most days, nobody is actively monitoring.

If we do this, the plan is to let a detection run for about 45 days without automatic isolation enabled to see if any false positives are caught.

Has anyone done this?  If so, did you regret it?  Or just business as usual?  Has it saved you yet?

document fetched 2026-09-06 00:33 · canonical https://reddit.com/r/sysadmin/comments/1w88ghv/do_you_automatically_isolate_serversdevices_based