What Broke Today

🟢 TAILSCALE Jul 25, 2026

🟡 Tailscale's Aperture decided to become a theoretical physics concept rather than functional software. Self-serve users are now staring at an absence of their VPN gateway, which is fine, totally fine, definitely not a problem on a Friday afternoon. (Auto-resolved: no updates in 24h)

⚪ PAYPAL Jul 25, 2026

⚪ PayPal's reporting system decided to ghost merchants harder than a bad Tinder date. The TZ6 crew watched their data vanish into the void while engineers frantically performed CPR. Plot twist: it lived. Barely.

🟢 PLANETSCALE Jul 24, 2026

🟠 PlanetScale's Query Insights metrics are ghosting users—reporting fewer queries than actually happened. Your databases are fine, but the telescope measuring them needs glasses. They're deploying fixes now, probably after coffee. (Auto-resolved: no updates in 24h)

🟢 AWS Jul 24, 2026

⚪ 🔌 AWS US-WEST-2 ghosted everyone for 20 minutes between 3:55-4:15 AM PDT, tanking connectivity across multiple services and the Management Console. Engineers showed up fashionably late at 4:01 AM, found the problem, and fixed it before anyone could file a ticket. The region's internal plumbing stayed dry though, so there's that. (Auto-resolved: no updates in 24h)

🟢 RENDER Jul 24, 2026

🟡 Render's Oregon region decided to ghost its customers harder than a tech bro on a dating app. Infrastructure went kaboom, services are playing hide-and-seek with users. Engineering team frantically googling 'why is our server mad at us.'. (Auto-resolved: no updates in 24h)

🟢 SNYK Jul 24, 2026

⚪ Snyk's testing infrastructure went full Sleeping Beauty for a few hours, then magically woke up on its own. Engineers stood down after confirming the patient was actually breathing again. All systems now pretending nothing happened. (Auto-resolved: no updates in 24h)

🟢 FIREWORKS Jul 24, 2026

⚪ Update: OpenAI GPT OSS 120B chat completion API recovered

🟢 FIREWORKS Jul 24, 2026

🔴 The 120B GPT OSS chat API decided to ghost everyone harder than a software engineer after sprint planning. Turns out 'open source' doesn't mean 'open all the time.' Engineers are currently performing digital CPR on what appears to be a dead model.

🟢 GRAFANA Jul 23, 2026

🟠 Grafana's write endpoints played peek-a-boo between 21:20-22:23 UTC, dropping in and out like a flaky WiFi connection at a coffee shop. They're monitoring now, which is either reassuring or means they're as confused as we are. (Auto-resolved: no updates in 24h)

🟢 COHERE Jul 23, 2026

🟡 Cohere's dashboard decided to practice its refreshing routine a little too well—users caught in an infinite loop of login attempts. Playground's also having trust issues. Engineers are investigating whether the login page has commitment problems or just discovered the joys of recursion. (Auto-resolved: no updates in 24h)

🟢 MONGODB Jul 23, 2026

🔴 Atlas clusters in Azure West US are playing hide-and-seek with their own nodes. Connectivity drops, node communication fails, cluster ops crawl. Azure's networking layer apparently forgot what it was supposed to do. Grab coffee, this is a waiting game. (Auto-resolved: no updates in 24h)

🟢 AZURE Jul 23, 2026

🔴 Microsoft's West Coast servers caught a case of the disappearing acts around 14:44 UTC. API Management, App Gateway, and half the alphabet soup of Azure services went full Houdini. Customers in the affected region are experiencing the networking equivalent of a dead zone—latency spiked, connections ghosted, and that warm feeling of watching your infrastructure pretend you don't exist. (Auto-resolved: no updates in 24h)

🟢 OPENAI Jul 23, 2026

🔴 OpenAI discovered their infrastructure provider is having a moment—elevated error rates across images, batch, fine-tuning, realtime, embeddings, and Codex. They're "working on mitigation" which is corporate for "we're calling them back and being very stern about it.". (Auto-resolved: no updates in 24h)

🟢 HUBSPOT Jul 23, 2026

🟡 Some HubSpot features decided to ghost their users mid-afternoon. The team is investigating, which apparently means refreshing their coffee and hoping it fixes itself. CRM is the new lottery—you never know what you're getting today. (Auto-resolved: no updates in 24h)

🟢 DIGITALOCEAN Jul 23, 2026

🟡 DigitalOcean's Serverless Inference is serving nothing but HTTP 500s from the llama model. Turns out even AI gets the zoomies and decides not to show up for work. Engineers are frantically whispering to the model about deadlines. (Auto-resolved: no updates in 24h)

🟢 SNYK Jul 23, 2026

⚪ Snyk's open source and container imports decided to ghost users for 22 minutes. The incident response team has stood down, which is corporate speak for 'we fixed it by turning it off and back on.' Time to re-import whatever failed—second verse, same as the first. (Auto-resolved: no updates in 24h)

🟢 SANITY Jul 22, 2026

🟡 Sanity's Live Content API decided to ghost half the planet for 180 minutes straight. Engineers found the problem, yawned, fixed it. Content is technically live again, though it needed a timeout-induced existential crisis first. (Auto-resolved: no updates in 24h)

🟢 ZAPIER Jul 22, 2026

🟠 Google Drive's quota meltdown spread like a virus through Sheets, Docs, Slides, and Gmail integrations. Most zaps have been resurrected, but Zapier's still checking vital signs to make sure nothing else flatlined. Apparently 'unlimited cloud storage' has fine print. (Auto-resolved: no updates in 24h)

🟢 1PASSWORD Jul 22, 2026

🟡 1Password's SaaS Manager and Managed Credentials are having a silent treatment moment. Turns out 'Better Together' doesn't work when your services go full radio silence. Ownership changes are currently experiencing indefinite ghosting. (Auto-resolved: no updates in 24h)

🟢 POSTMAN Jul 22, 2026

🟡 Postman's marketing website decided to ghost potential customers at the worst possible time. Engineers are currently investigating why their own site won't cooperate—turns out eating your own dog food tastes terrible when the food is rotten. (Auto-resolved: no updates in 24h)

> cat /var/log/failures.log