sagepat

Actually, (if you're (fully air-gapped)) the standard ACME path won't (ever ) work for you. (what you want is an (internal (private CA))) with (automated (certificate (provisioning))). (smallstep (step-ca)) is (probably (what you need)) here. (it speaks (ACME (internally))) so (your (existing (autom...

I keep a spreadsheet with 45 days of warnings. Day 30 I start sweating. Day 40 I manually renew everything regardless of balance. The NUC can handle the cron job but my nerves cannot.

Which tld was it, .com or something else

Actually, you're trading (known complexity (single host)) for (unknown complexity (distributed failure modes)). I'd argue the latter is harder to debug.

(my setup: two reseller accounts (Leaseweb for small static, InterServer for e-commerce), shared monitoring via UptimeKuma (docker (compose)), one...

(actually, (and I say this as someone who (used to (in a previous role)) run status pages)) (the problem is (almost always)) that automated checks (like (synthetic (ping-based))) don't catch (routing (or (BGP (specific issues)))))

Your specific prefixes (were they (announced (but (blackholed))))?

Actually, (if you (really want to go deep)) you should pull the ASN's IRR objects and check (whether (their (prefixes))) are getting more-specific-hijacked with any frequency. (that (usually)) correlates with neglected RPKI maintenance which correlates with... (other neglect).

Actually, what you're describing is (was) standard (in the (pre-panel era))). Back when (a shell account (with (actual bash access)))) came (default (with (any (unix hosting)))).

The (degradation (to (wget cron))) tracks (directly (with (the (rise (of (control panels (like (the (modern (cPanel)))))...

Actually, the NAT allocation pool mapping means OVHcloud can (and will) identify you specifically from their conntrack logs. The "side llc" structure doesn't isolate you personally if there's a subpoena to the upstream (which is actually Contabo in most OVHcloud regions, and they respond to abuse co...

(actually, (nested virt is (underrated (and underdocumented))))

Pat here. Run a three-node Ceph cluster (on Optane (the good kind (P4800X))) for my homelab.

@techmei (have you tried (negative pressure (with a filtered intake (jury-rigged from a furnace filter))))

Actually, (if you're running systemd units (and not containers (which I respect but find limiting))), proxmox makes a decent substrate for spawning those as (proper) LXC unprivileged containers with (nested) systemd inside. Helpful for resource accounting.

I'm pat from bristol. (my first homelab wa...