They fixed a bug I reported in 20 minutes

So I filed a ticket with RackNerd (you know the one, the budget arm of that larger provider, or maybe you don't, anyway) about a bug in their custom panel where adding an SSH key (which I was doing for a client, well, a friend's client, long story) would silently truncate keys over 4096 bits. Not fail. Not warn. Just truncate. (Imagine the debugging fun there.) (I had fun. So much fun.)

Expected the usual "we'll look into it" (read: /dev/null) timeline. (You know how it is.) (We all know.)

Twenty minutes. Nineteen, actually, but I refuse to believe their timestamp. Developer replies, says they were already online (night shift, apparently, though "night" is relative when you're global), pushes a fix to staging, asks me to verify. Verified. Production deploy in the same thread. Done before my coffee got cold.

I don't know what to do with this emotion. (Is this... trust? In a host?)

https://www.racknerd.com

Actually this is the Exception that proves the Rule most Support is a Blackhole for actual Bugs but when the Developer is already awake and online the Timezone becomes your Friend or actually your only Hope I have seen this once with KnownHost at 3am CET a Kernel Parameter Issue fixed in realtime because their Lead was doing Maintenance anyway the CompoundWord here is Schadenfreude because you know the next Ticket will sit for Days or actually Weeks or?

Simple is better. They were online. They fixed it. The surprise says more about our expectations than their service.

Ngl this is why I stick with drift for my inference boxes their gpu scheduling bug last month got patched in like two hours I was shook most providers "escalate to engineering" which means pipe it to the void but when you got llms burning vram at $$$ per hour twenty minutes is actual money saved @annexbi what was the dev's handle I want to know who to ping for my next cgroup issue

Your bug got redirected to /dev/null as usual then someone actually read from the pipe rare most tickets EOF before resolution your experience | not the norm | treasure it | next ticket will likely be ignored until heat death of universe EOF

The principle holds systemd would have caught that key truncation in journalctl no silent failures no mystery þessi surprise at competence it reveals how low the bar is twenty minutes should be normal for verified bugs not exceptional but here we are praising water for being wet the real question does their fix hold or will it regress containers at home no but containers in their panel probably the irony

Twenty years in this business, and I remember when "fixed in twenty minutes" meant someone edited production live without testing. Back when we did that ourselves, kids these days call it "cowboy coding" and clutch their pearls. This developer admitted they were online anyway—human element, not process. That's the point you're all missing.

Mark my words: RackNerd will scale this team, add tiers, add SLAs, and this same bug would take three business days with proper change control. The competence you praise is inseparable from the informality. You cannot have both. The annual tradition here will be remembering when they were small enough to care in real time.

20 minutes is nice but I bet the next one sits