What Broke Today

🔴 GITHUB 2h ago

🟠 GitHub decided to ghost developers twice in one day—Git ops, Issues, Actions, and PRs all took an unscheduled siesta from 16:52-17:01 UTC. Systems are back online, but the incident playlist is starting to feel like a broken record.

🔴 LOVABLE 5h ago

🟡 Lovable's editor decided messages are optional now. Users report errors when trying to send anything—turns out the platform's commitment to communication was mostly aspirational. Engineers are speedrunning a fix before the irony becomes weaponized.

🔴 OPENAI 13h ago

🟡 OpenAI's desktop app decided Work threads are overrated. Users now experiencing the cutting-edge feature of threads that simply don't exist. Engineering team is 'investigating' what surely won't take 6 hours to fix.

🔴 OPENAI 13h ago

🟡 OpenAI's desktop app decided threads were too mainstream, now refusing to create them for regular users. Codex is having commitment issues. Engineers are investigating whether this was a feature or a bug—spoiler: it's a bug.

🔴 KRAKEN 15h ago

🟡 Krak Card API decided to ghost European and UK users right when they needed it most. Transactions and orders are failing faster than crypto during a bear market. Engineers are frantically checking if they paid the API's electricity bill.

🔴 SNYK 19h ago

⚪ Snyk's dependency scanner got too ambitious analyzing test failures and face-planted hard. Engineers threw it in recovery mode, got it back on its feet, and confirmed the patient is breathing normally again. Crisis averted—for now.

🟢 SENTRY 1d ago

🟡 Sentry's dashboard went blind in both US and EU regions. The irony of an error tracking platform becoming untrackable is *chef's kiss*. Engineers are investigating whether they need to file a bug report against themselves. (Auto-resolved: no updates in 24h)

🟢 SNYK 1d ago

🟠 Snyk's vulnerability scanner discovered its own weakness: gateway timeouts turning 504s into a customer experience audit. Engineers found the problem and are speed-running fixes while tests crawl at 45-second intervals—proof that even security tools need security blankets. (Auto-resolved: no updates in 24h)

🟢 ELEVENLABS 1d ago

🟡 ElevenLabs' EU telephony agents decided to ghost SIP calls entirely. Turns out 'voice AI' doesn't work when the voice part breaks. Engineering team is currently using Ouija board to communicate with the infrastructure. (Auto-resolved: no updates in 24h)

🟢 ZAPIER 1d ago

🟡 Zapier's Facebook and Microsoft login buttons led to the void after authentication. Engineers found the culprit and shipped a fix, now playing the waiting game to see if it actually stuck or just created new and exciting problems. (Auto-resolved: no updates in 24h)

🟢 AMPLITUDE 2d ago

⚪ Amplitude's Guides and Surveys service took an unexpected 8-minute nap between 11:40-11:48 AM PT, leaving users unable to create, edit, or publish anything. They've since woken up and claim everything's fine, though they're still checking if anyone left the stove on. (Auto-resolved: no updates in 24h)

🟢 CLOUDINARY 2d ago

⚪ Your image metrics ghosted you for 33 minutes while Cloudinary's dashboard decided to stop counting. The good news: they found them under the couch. Everything's back to pretending to track your bandwidth now. (Auto-resolved: no updates in 24h)

🟢 CODECOV 2d ago

{ "skip": false, "headline": "Codecov vanishes into the void, investigation ongoing", "summary": "🔴 Codecov's entire app and services decided to ghost everyone simultaneously. Engineers are "investigating" which is corporate for "frantically searching the logs with a flashlight." Your code coverage reports are taking an unscheduled vacation." }. (Auto-resolved: no updates in 24h)

🟢 DIGITALOCEAN 4d ago

⚪ BLR1 region went full Sleeping Beauty from 11:59-13:02 UTC, ghosting Monitoring, App Platform, Databases, Functions, Load Balancers, Spaces, and the control panel. Users got the classic combo of latency, dropped connections, and errors—basically a networking horror movie where the power came back before anyone got to the climax. (Auto-resolved: no updates in 24h)

🟢 FLY 4d ago

🟠 Fly's Amsterdam region decided to take an unscheduled nap. Management plane and AMS hosts are stumbling back online while engineers frantically assess the damage. MPG impact assessment in progress—translation: they're still counting the bodies. (Auto-resolved: no updates in 24h)

🟢 LANGSMITH 4d ago

🟡 LangSmith deployments are choking during database setup across GCP regions—turns out their cloud provider's infrastructure decided to take an unscheduled nap. The Control Plane is having an identity crisis and can't spin up new dedicated instances. Someone's definitely stress-testing their oncall rotation right now. (Auto-resolved: no updates in 24h)

🟢 OPENAI 6d ago

🟡 OpenAI's login, signup, and ads infrastructure caught a sudden case of amnesia. Users can't get in, ads won't serve, and the API is playing dead. Nothing says 'we have this under control' like simultaneously breaking the front door and the cash register. (Auto-resolved: no updates in 24h)

🟢 OPENAI 6d ago

🟡 OpenAI's login system caught in an identity crisis—you can't get in, they can't get in, nobody gets in. The irony of an AI company unable to authenticate humans is *chef's kiss*. (Auto-resolved: no updates in 24h)

🟢 PYPI 6d ago

🔴 Multiple python.org services went dark after deprecated API usage tanked their internal certificates. The classic 'we updated the library and broke everything' play. Now frantically renewing certs and praying things resurrect. (Auto-resolved: no updates in 24h)

🟢 ELEVENLABS Sep 29, 2026

🟠 ElevenLabs' speech-to-text took a vow of silence while Agents and Creative Platform discovered what 'latency' really means. Engineers found the culprit but apparently need a moment to perform CPR on their infrastructure. (Auto-resolved: no updates in 24h)

> cat /var/log/failures.log