So Thursday afternoon I'm sitting at my desk trying to queue up a podcast before a call and Spotify just... stops. Spinning wheel, then a grey error screen, then nothing. My first move, like an idiot, was to restart my router. Ten minutes of that before I finally checked Twitter (sorry, X) and realized it wasn't me. It was basically everyone.
That was June 12th. A chunk of Google Cloud fell over that morning, something in their service management layer, an API config change that went sideways and cascaded outward, and because half the internet quietly runs on the same three or four backbones now, the blast radius was enormous. Spotify was down. Discord was down. Snapchat wouldn't load. Even Cloudflare had a rough go of it later in the day, which is its own funny irony since Cloudflare is the thing a lot of sites lean on specifically so a single point of failure somewhere else doesn't take them out.
I don't want to pretend I have some deep technical postmortem here, because I don't work in infrastructure and I'm not going to fake expertise I don't have. What I do have is a very ordinary Thursday where three separate apps I use every day just quietly died within about twenty minutes of each other, and that's the part that stuck with me more than the actual cause. We've built this weird situation where "the cloud" sounds like it means everywhere and nowhere, distributed, resilient, but in practice a lot of it funnels through a small number of companies' data centers, and when one of those companies has a bad morning, you feel it in your headphones.
I keep about 40 gigs of MP3s on an old external drive specifically because of days like this, and I felt smug for about four minutes before I remembered the drive was at my parents' place from when I helped set up their smart TV in April. So that plan failed too. Ended up just sitting there in silence for my call, which honestly was fine, but it's a little embarrassing how dependent I've gotten on a streaming service to provide background noise for a Zoom call I wasn't even talking much in.
The other thing that got me was how long it took some of these companies to actually say anything useful. Status pages are their own genre of corporate non-writing — "we are investigating elevated error rates" is doing a lot of work to avoid saying "everything is broken and we don't know why yet." I get why they write like that, legal probably has notes, but as a user sitting there refreshing Downdetector every ninety seconds it's not exactly reassuring.
Worth saying too: this happened the same week as Apple's WWDC keynote, which ate up basically all the tech coverage oxygen, and everyone was busy arguing about the new "Liquid Glass" look and whether jumping straight from iOS 18 to iOS 26 was a smart naming move or just marketing trying to match the model year on your car. Fair discussion to have, I guess, though I'll admit I rolled my eyes a little at yet another visual refresh being treated like the biggest tech story of the month. Meanwhile a genuinely significant outage that knocked out chunks of the consumer internet for a few hours got maybe a day of coverage before everyone moved back to arguing about translucent icons. Priorities.
By early evening most things were back. Spotify came back for me around 4:40pm my time, Discord a bit before that. No apology email, no "here's what we're doing differently," just... it worked again, so we all moved on like nothing happened. Which I guess is the normal cycle at this point. These things happen every few months now, somewhere in the stack, and we've all gotten weirdly good at just waiting it out instead of being alarmed by it.
Anyway. My external drive is coming back from my parents' place this weekend and it's staying on my desk from now on, outage or no outage. Small lesson learned, nothing profound about it.