> cat /var/log/failures.log
A comprehensive record of when things went wrong.
๐ด SNYK
1h ago
โช Snyk's authentication decided to play peek-a-boo with users, then got fixed faster than you can say 'dependency vulnerability.' The incident response squad has clocked out, services are stable, and your secrets are marginally safer again.
๐ด HASHICORP
7h ago
๐ก The service that manages your infrastructure is too busy managing its own meltdown. Errors and latency spiking across runs and UI while engineers play technical detective with the status page refresh button.
๐ด GITHUB
14h ago
๐ Copilot's MAI Code 2 Flash decided to ghost developers harder than your ex. GitHub's been playing tech support therapist with their upstream provider and things are allegedly 'improving' now. Translation: they hit it with a wrench and some prayers.
๐ข SNYK
1d ago
๐ก Snyk's app management features went into hibernation while existing apps just sit there, perfectly fine, watching the chaos. Engineers are currently having a staring contest with their monitors trying to figure out what happened. (Auto-resolved: no updates in 24h)
๐ข PLANETSCALE
1d ago
๐ก PlanetScale's Postgres databases in Azure decided to ghost users across Europe, Iowa, and Virginia simultaneously. Investigators are searching for them, though at this point they're probably just hiding under a desk somewhere hoping nobody finds them. (Auto-resolved: no updates in 24h)
๐ข AWS
3d ago
โช AWS's identity service decided to play gatekeeper during peak hours, locking out auth-dependent services across the region. Three-hour lockout from 3-6 AM PST hit IAM Identity Center, Connect, and anything else that actually needs login credentials to function. The post-mortem will definitely claim 'cascading network events' instead of 'we broke authentication.'. (Auto-resolved: no updates in 24h)
๐ข HASHICORP
4d ago
๐ก Terraform's private infrastructure is throwing 403s at everyone like a bouncer who lost the guest list. HYOK, private VCS, and run tasks are getting the cold shoulder through tfc-agent. Engineers are now playing detective with what should've been a rubber stamp. (Auto-resolved: no updates in 24h)
๐ข WORKOS
5d ago
โช WorkOS's single sign-on decided to live up to its name by making users sign in... never. From 09:48-09:53 UTC, SSO auth crawled, identity callbacks ghosted, and token exchange discovered the concept of 'timeout.' Five minutes later, whatever deity they prayed to answered. (Auto-resolved: no updates in 24h)
๐ข GITHUB
5d ago
๐ Pull requests decided to move at geological timescalesโ3% of merges hit the timeout wall. GitHub applied the digital equivalent of smacking it on the side, knocked it down to <1%. Still holding its breath though. (Auto-resolved: no updates in 24h)
๐ข ANTHROPIC
5d ago
๐ Anthropic's golden goose Claude.ai, the API, and Code all decided simultaneous error rates were trending. The connectors ghosted users but are slowly sliding back into DMs. Adding new connectors? Still in timeout jail. Scheduled tasks are fashionably lateโlike a two-hour "I'll be there in five minutes" late. (Auto-resolved: no updates in 24h)
๐ข LINEAR
5d ago
๐ Linear's MCP connector decided geography doesn't matter when playing dead. Both US and EU regions watching a downstream provider implode like a house of cards. Engineering team is monitoring, which is corporate speak for 'we're waiting to see if it fixes itself.'. (Auto-resolved: no updates in 24h)
๐ข NETLIFY
5d ago
๐ Microsoft's Exchange Online decided support tickets deserved the slow treatment, so Netlify's responses are taking the scenic route to Outlook inboxes. Support team is working, but your emails are stuck in Microsoft's spam purgatory. (Auto-resolved: no updates in 24h)
๐ข SUPABASE
6d ago
๐ก One customer's API went full Niagara Falls and Supabase initially thought the sky was falling. Turns out it was just a single project hammering endpoints like a toddler discovering a doorbell. Everything's actually fine now, but that was a fun 10 minutes of false alarm theater. (Auto-resolved: no updates in 24h)
โช WEBFLOW
6d ago
โช Webflow's hosted sites decided to ghost visitors for 13 minutes straight. The good news: it's resolved. The bad news: your customers experienced a masterclass in 404 existentialism. Crisis averted, dignity questionable.
๐ข MONDAY
6d ago
๐ก Monday's Outlook integrations became sentient and decided to ghost everyone. Turns out when Microsoft sneezes, Monday catches pneumonia. Their team is now playing 4D chess with the outage, hoping to restore sanity before users storm the forums. (Auto-resolved: no updates in 24h)
๐ข SUPABASE
6d ago
๐ Supabase's connection pooling decided regions sa-east-1 and ap-southeast-1 weren't worth talking to anymore. Some databases are playing dead while engineers frantically check if they remembered to pay the infrastructure bill. (Auto-resolved: no updates in 24h)
๐ข SNYK
6d ago
โช Snyk's Go-based PR checks went full existential crisis, then got therapy. Engineers applied the digital equivalent of duct tape, success rates stabilized, and everyone pretended this never happened. Crisis averted, dignity remains questionable. (Auto-resolved: no updates in 24h)
๐ข PAYPAL
6d ago
โช PayPal's payment APIs decided to collectively ghost every transaction type simultaneouslyโrefunds, authorizations, checkouts, the works. Engineers found the off switch and flipped it back. Global catastrophe, now quietly resolved with zero explanation of what actually broke. (Auto-resolved: no updates in 24h)
๐ข BUNNY_CDN
6d ago
๐ Turns out LA's power grid doesn't have a failover. Bunny's LA infrastructure took an unscheduled nap, forcing CDN traffic into emergency relocation. Some servers are back online; others are still meditating on their mortality. (Auto-resolved: no updates in 24h)
๐ข FLY
6d ago
๐ Fly's Los Angeles region decided to experience a power outage because apparently the grid runs on hopes and prayers. Apps and Postgres clusters are now as responsive as a broken answering machine. Upstream datacenter forgot to pay the electric bill. (Auto-resolved: no updates in 24h)
Page 1 of 55
Older disasters โ