Is Recent System Drift Causing Multiple Silent Failures?

DKIM is not necessarily evidence of a kernel problem. Sending services, selectors, SPF includes, subdomains, and DMARC alignment can drift when someone adds an email provider or makes a temporary configuration change. Periodically verify the published records and selectors against the systems that are actually sending mail. For the host issues, I’d still inspect package histories, hardware health, memory errors, disk space, and the vendor’s kernel configuration before blaming upstream kernel changes.

The cron logging issue could be caused by log rotation or a size and file-count limit rather than cron itself. Some logging setups stop behaving once a directory accumulates too many files or reaches an unexpected limit. Check the rotation rules, disk usage, inode usage, and whether the job is still running but writing somewhere else.

I’d compare kernel and package versions from before the problems started instead of treating every failure as unrelated. Several different issues appearing in the same time window makes the package update history worth investigating, even if no one changed anything manually. Automatic updates are still changes.

The bigger pattern is that everything failed silently. Add external checks based on outcomes rather than just process status: alert before certificates expire, use a heartbeat or dead-man’s switch for cron, monitor container restart and OOM counts, and test the actual WireGuard tunnel. That makes drift visible soon after it happens.