> cat /var/log/failures.log
A comprehensive record of when things went wrong.
🟢 GRAFANA
Jul 15, 2026
🟠 Grafana's prod-eu-west-2 decided to do the Schrödinger's outage thing—some dashboards work, others don't, nobody knows which. Third-party provider apparently forgot to pay their bills. Investigators are doing that classic 'we're looking into it' dance while engineers stress-refresh Slack. (Auto-resolved: no updates in 24h)
🟢 MUX
Jul 15, 2026
🔴 Video ingestion and playback went full Thanos snap across one data center. Mux rerouted traffic like they were directing lost tourists, now playing blame ping-pong with their cloud provider. (Auto-resolved: no updates in 24h)
🟢 GCP
Jul 15, 2026
⚪ Update: RESOLVED: Google Cloud VMware Engine (GCVE) Stretched Cluster customers are experiencing zonal outag
⚪ WEBFLOW
Jul 15, 2026
⚪ Webflow locked customers out of their accounts for about an hour overnight. The good news: they fixed it. The bad news: your auth system chose violence at 11:30 PM UTC. At least the incident report will be shorter than your rebuild time.
⚪ WEBFLOW
Jul 15, 2026
🟠 Webflow decided to play 'are you who you think you are?' with login credentials. They've apparently stopped the bleeding, but are staying glued to their monitors like anxious parents waiting for test results. Mitigation deployed, fingers crossed.
🟢 RENDER
Jul 15, 2026
🟡 Render's dashboard decided to ghost its users mid-July. Engineers are currently performing archaeological digs through logs, muttering about 'availability issues' like it's not their own product. You can still deploy things, but good luck knowing if it worked. (Auto-resolved: no updates in 24h)
🟢 OPENAI
Jul 15, 2026
🟡 OpenAI is 'investigating' why basically every ChatGPT feature simultaneously decided to ghost users. Search, conversations, agents, Atlas, sites, research, connectors, and voice all throwing errors like they coordinated a walkout. Even the desktop app's code generation is taking an unscheduled sabbatical. (Auto-resolved: no updates in 24h)
🟢 ANTHROPIC
Jul 14, 2026
🟡 Anthropic's investigating why Claude suddenly forgot how to build containers. Document creation, Cowork Remote, and Claude Code Remote are all experiencing existential crises. Your AI's having an identity crisis—turns out even genius needs a working dev environment. (Auto-resolved: no updates in 24h)
🟢 GCP
Jul 14, 2026
🔴 Google's VMware Engine decided zones weren't worth talking to each other. Multi-region connectivity is playing dead while engineers dust off the VMware manual they swore they'd never need. Stretched clusters are now just... unstretched.
🟢 GRAFANA
Jul 14, 2026
🟠 Grafana's metrics backend decided partial outages were the new feature release. Writes are limping along, reads are confused, and rule evaluation is just vibing offline. Engineers are performing emergency CPR in the datacenter. (Auto-resolved: no updates in 24h)
🟢 FLY
Jul 14, 2026
🟠 Fly's San Jose region decided some hosts needed an unscheduled vacation. Your app in SJC? Might be reachable, might not—it's a lucky dip. Engineers are currently checking if the servers are actually offline or just ignoring Slack messages. (Auto-resolved: no updates in 24h)
🟢 AMPLITUDE
Jul 13, 2026
🟠 Amplitude's UI and MCP APIs decided to ghost their users since 2:04 AM PT. Team's investigating while your dashboards contemplate the void. Data supposedly intact, but who's counting anymore?. (Auto-resolved: no updates in 24h)
🟢 MAILGUN
Jul 13, 2026
🔴 Mailgun's US email validation is throwing 500s like confetti and the control panel decided to play hide-and-seek. Turns out validation is hard when your servers are having an existential crisis. (Auto-resolved: no updates in 24h)
🟢 MAILGUN
Jul 13, 2026
🟡 Mailgun's validators started throwing existential 500s while the control panel decided it was done being controllable. Some users can access it, others get the digital shoulder shrug. Schrodinger's email service. (Auto-resolved: no updates in 24h)
🟢 1PASSWORD
Jul 13, 2026
⚪ 1Password's update server decided to ghosted GitHub Actions, turning deploy pipelines into archaeological digs. The fix is supposedly in place, but the team's still glued to their monitors like it owes them money. (Auto-resolved: no updates in 24h)
🟢 AMPLITUDE
Jul 13, 2026
🟡 Amplitude's Custom Metrics details page is playing hide-and-seek with users—loads fail, errors appear, data supposedly safe. Engineering team is *actively investigating*, which in status-page speak means they're still figuring out which Slack thread has the answer. (Auto-resolved: no updates in 24h)
🟢 CLERK
Jul 12, 2026
🟠 Clerk's domain configuration decided to cosplay as a broken record—elevated errors while provisioning means your 'Go live' button is more 'Go cry' right now. Engineers are apparently monitoring the disaster unfold in real-time. (Auto-resolved: no updates in 24h)
🟢 OPENAI
Jul 11, 2026
🟡 OpenAI's text-to-video generator has achieved perfect invisibility—turns out the real special effect was becoming unavailable. Engineers are currently investigating why their generative AI can't generate... anything. The irony is so thick you could render it. (Auto-resolved: no updates in 24h)
🟢 SENTRY
Jul 11, 2026
🟠 Sentry's US region went full ostrich mode on error ingestion, now playing archaeological dig to recover 25M lost events. They keep saying '2 hours' with the confidence of a contractor giving time estimates. (Auto-resolved: no updates in 24h)
🟢 OPENAI
Jul 11, 2026
⚪ OpenAI's FedRAMP customers discovered that Codex, analytics, search, custom GPTs, invites, and compliance downloads had gone full Thanos. Turns out they just needed a nap. All features have been resurrected and are now asking existential questions about what happened. (Auto-resolved: no updates in 24h)