how to troubleshoot spurious network timeouts

We have a "Apple Studio M1" with macOS 13.6.1 (Ventura) in one of our studios. As the machine is little used but running 24/7, I've set it up as a CI-runner for our local GitLab instance (using gitlab-runner+tart). Mostly things work as expected. However, every now and then, the machine becomes unreachable over the network. That is: - I **can** still ping the machine - trying to ssh into the machine stalls - open ssh sessions become unresponsive after a while (a couple of minutes), the system goes back to normal, and I can resume my work in the ssh session (or login via ssh again). the gitlab-ci jobs are run within a virtual machine (tart is using the *Virtualization.Framework*). if a gitlab-ci job has been running during the network outage, it appears to keep running fine. only, if it needs to access the network itself (e.g. because it installs something via brew, or tries to download stuff otherwise,...) the job will fail with something like this:

$ which automake || brew install automake
automake not found
==> Fetching dependencies for automake: m4 and autoconf
==> Fetching m4
==> Downloading https://raw.githubusercontent.com/Homebrew/homebrew-core/c8215d5f7473269cd331dcd8cd3ab38e7c26dfe9/Formula/m/m4.rb 
curl: (6) Could not resolve host: raw.githubusercontent.com

the host has currently some problems with its monitor (it's behind an oldish KVM switch) and besides it is in another floor, so its impractical to check the situation directly at the machine. on the plus side, this means that I am pretty sure that nobody is fuddling with the settings :-) as i'm a linux guy by trade, i would like some advice on how to debug this issue (e.g. which logs to check, which tools to run,...) my naive checks of some of the things like system.log revealed nothing. e.g. i logged into the machine about an hour ago, and the connection stalled at about 11:45. it is now (12:05) reachable again, and the syslog says:

Jul 29 09:07:30 macstudio syslogd: ASL Sender Statistics
Jul 29 09:17:30 macstudio syslogd: ASL Sender Statistics
Jul 29 09:27:30 macstudio syslogd: ASL Sender Statistics
Jul 29 09:37:30 macstudio syslogd: ASL Sender Statistics
Jul 29 09:47:30 macstudio syslogd: ASL Sender Statistics
Jul 29 09:57:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:07:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:17:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:27:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:37:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:47:30 macstudio syslogd: ASL Sender Statistics
Jul 29 10:57:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:07:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:17:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:27:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:37:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:39:02 macstudio sshd: macuser [priv]: USER_PROCESS: 97643 ttys000
Jul 29 11:47:30 macstudio syslogd: ASL Sender Statistics
Jul 29 11:56:34 macstudio sshd: macuser [priv]: USER_PROCESS: 97677 ttys001
Jul 29 11:57:16 macstudio sshd: macuser [priv]: USER_PROCESS: 97677 ttys005
Jul 29 11:58:19 macstudio syslogd: ASL Sender Statistics
Jul 29 12:02:19 macstudio sshd: macuser [priv]: USER_PROCESS: 97677 ttys006

(so to my eye there is no trace). the install.log has an interesting entry at about the time the host came back: "System has powered on":

2025-07-29 08:32:31+02 macstudio SoftwareUpdateNotificationManager: AssertionMgr: Cancel com.apple.softwareupdate.NotifyAgentAssertion-UpdatesAvailable assertion for pid 587, id 0x8B7B
2025-07-29 08:32:31+02 macstudio SoftwareUpdateNotificationManager: AssertionMgr: Cancel com.apple.softwareupdate.NotifyAgentAssertion-UpdatesAvailable assertion for pid 10548, id 0x8B7A
2025-07-29 12:02:25+02 macstudio SoftwareUpdateNotificationManager: SUOSUPowerEventObserver: System will power on
2025-07-29 12:02:25+02 macstudio SoftwareUpdateNotificationManager: SUOSUPowerEventObserver: System has powered on
2025-07-29 12:02:25+02 macstudio softwareupdated: SUOSUPowerEventObserver: System will power on
2025-07-29 12:02:25+02 macstudio softwareupdated: SUOSUPowerEventObserver: System has powered on
2025-07-29 12:07:50+02 macstudio SoftwareUpdateNotificationManager: SUOSUPowerEventObserver: System is considering sleep
2025-07-29 12:07:50+02 macstudio softwareupdated: SUOSUPowerEventObserver: System is considering sleep