> cat /var/log/failures.log
A comprehensive record of when things went wrong.
🟢 SNOWFLAKE
Jun 22, 2026
🔴 Snowflake in US West decided to play hide-and-seek with customer data. Engineers are investigating why 'processing efficiency' is code for 'we broke it.' Meanwhile, your queries are getting real acquainted with timeout errors. (Auto-resolved: no updates in 24h)
🟢 LINEAR
Jun 22, 2026
🟡 Linear is having a staring contest with your data and winning. US customers getting the spinning wheel of contemplation while engineers forensically examine what went wrong. Update promised by 17:45 UTC, which in status-page time means 'we're still Googling'. (Auto-resolved: no updates in 24h)
🟢 SNYK
Jun 22, 2026
{
"skip": true
}
```
This is a resolved incident notification (issue is already fixed and back to normal), so it gets skipped per the guidelines for scheduled/resolved maintenance items. (Auto-resolved: no updates in 24h)
🟢 JIRA
Jun 19, 2026
🟠 Jira's email-to-ticket pipeline forgot how to read Microsoft OAuth credentials, leaving some customers with a pile of unprocessed messages. Now fixed, but affected folks need to re-auth their mail servers like it's their first day on the internet again. (Auto-resolved: no updates in 24h)
🟢 VERCEL
Jun 19, 2026
⚪ 🔧 Vercel's Functions threw ERR_STREAM_PREMATURE_CLOSE errors after a Node.js security patch introduced a regression that murdered response streaming. node-fetch@2 users got the short end of the stick for ~7 hours while engineers played upstream whack-a-mole. (Auto-resolved: no updates in 24h)
🟢 GRAFANA
Jun 19, 2026
🔴 Your alerts were screaming but nobody heard them. Grafana's prod-us-central-0 region had a brief existential crisis with rule evaluations (15:16-15:59 UTC), which is just the worst time for monitoring to stop monitoring. Engineering found the problem, patched it, and now they're nervously refreshing dashboards. (Auto-resolved: no updates in 24h)
🟢 DIGITALOCEAN
Jun 19, 2026
🟡 DigitalOcean's graphs decided to practice invisibility—monitoring data exists, it's just playing hide-and-seek in the control panel. Droplets, load balancers, databases all flying blind. Engineering team frantically checking if they accidentally deleted the entire 'show stuff' feature. (Auto-resolved: no updates in 24h)
🟢 UPSTASH
Jun 19, 2026
⚪ QStash EU region went full hermit mode for 34 minutes—some requests got ghosted during the networking blackout. They've since remembered that connectivity exists and are back to normal service. Postmortem TBA, presumably titled 'How We Lost Our Internet Privileges.'. (Auto-resolved: no updates in 24h)
🟢 CONTENTFUL
Jun 18, 2026
🟠 Contentful's Management and Preview APIs decided to ghost some customers with 500 errors, though most folks are blissfully unaware their content delivery infrastructure is having a localized meltdown. Engineers found the culprit lurking in specific CDN nodes and are currently performing digital archaeology to fix it. (Auto-resolved: no updates in 24h)
🟢 GRAFANA
Jun 18, 2026
🟡 Grafana's having a geography problem: users in India can't load dashboards while Google's datacenter decides to take an unscheduled vacation. Asset loading errors everywhere. GCP's chaos, Grafana's reputation damage. (Auto-resolved: no updates in 24h)
🟢 GRAFANA
Jun 18, 2026
⚪ Grafana's k6 load testing UI had a brief identity crisis where it forgot what 'performance' meant. Engineers applied the digital equivalent of smelling salts and it's back to normal. Lesson: even tools that test performance can't test themselves apparently. (Auto-resolved: no updates in 24h)
🟢 SNYK
Jun 18, 2026
⚪ Snyk's web portal took an unscheduled nap affecting imports, retests, and PR checks. Engineers poked it with a stick, it woke up, and pretended nothing happened. Incident response stood down at 10:59 UTC. (Auto-resolved: no updates in 24h)
🟢 ANTHROPIC
Jun 18, 2026
🔴 Anthropic's flagship AI decided to ghost its users while engineers play detective with the source code. 'We identified the problem' is corporate for 'we found where it broke, now comes the fun part.' ETA on consciousness restoration: TBD. (Auto-resolved: no updates in 24h)
🟢 SUPABASE
Jun 18, 2026
🔴 DNS record creation decided to return 500 errors like a petulant teenager, tanking project creation, pausing, and unpausing. Supabase hit the emergency brake and disabled new projects while engineers pray to the DNS gods. Impact still cascading through their infrastructure like a game of Jenga played by a toddler. (Auto-resolved: no updates in 24h)
🟢 SUPABASE
Jun 18, 2026
🟡 Supabase's DNS API decided to ghost new project creation attempts. Engineers are frantically checking if they unplugged the right cable this time. The irony of a backend service making backends inaccessible is not lost. (Auto-resolved: no updates in 24h)
🟢 LINEAR
Jun 17, 2026
🟠 Linear decided to test how slowly it could move tasks around. After a rollback that'll definitely haunt someone's commit history, performance is limping back to normal. Integrations are also getting their act together. (Auto-resolved: no updates in 24h)
🟢 LAUNCHDARKLY
Jun 17, 2026
🟠 LaunchDarkly.com and docs decided to play hide-and-seek with users while the actual flag delivery service flexed its uptime muscles. At least the part that matters still works—the website can take a permanent vacation for all anyone cares. (Auto-resolved: no updates in 24h)
🟢 OPENAI
Jun 17, 2026
🟠 OpenAI's mobile users discovered their conversations were having an existential crisis. Engineers applied a band-aid and are now watching to see if it sticks. Turns out 'Conversations' needed to remember how to actually converse. (Auto-resolved: no updates in 24h)
🟢 OPENAI
Jun 17, 2026
🟠 OpenAI's Conversations feature is throwing errors on mobile like a teenager during family dinner. They've 'identified' the problem, which in tech means they found the smoking gun but the shooter's still at large. Mitigation incoming, fingers crossed. (Auto-resolved: no updates in 24h)
🟢 SUPABASE
Jun 17, 2026
⚪ Supabase sweated through a hardware failure in us-east-2 that left projects wondering if they were dumped. The team resurrected all affected databases by 17:02 UTC. If your project's still not texting back, they want to hear from you. (Auto-resolved: no updates in 24h)