No widespread internet outage was detected today. Several smaller, product-specific incidents are active.
Active incidents
Cloudflare: elevated errors for D1 (500 and error 7500 responses) and for Workers Workflows, both opened around 09:22-09:24 UTC and still under investigation. Workflows success rates have been dipping since about 07:00 UTC. These affect specific products, not Cloudflare's CDN as a whole. cloudflarestatus.comTwilio: several carrier-specific issues, including Lookup API failures on Verizon (US), call failures to T-Mobile (US, recovery being monitored), SMS delays and failures in Brazil and to Docomo Pacific, and voice post-dial delay to Thailand. status.twilio.comAkamai: a fix for edge delivery issues between Indonesia and Singapore was applied Oct 25 and is still being monitored. A brief packet-loss episode affecting Mumbai and Chennai (08:05-10:45 UTC, upstream provider) has subsided but is not yet marked resolved. Akamai is also assessing two security advisories (Brotli and F5/NGINX vulnerabilities). akamaistatus.comRecently resolved (past 72 hours)
DigitalOcean: Managed Kubernetes cluster operations were delayed or failing across regions (Oct 27 22:14 UTC to Oct 28 01:55 UTC). Existing workloads were not affected.GitHub: degraded availability of some Copilot models on Oct 28, tied to upstream model providers. Resolved.Anthropic: Google Calendar and Gmail connector errors in claude.ai and Claude Code (Oct 26-28). Resolved.All clear at the time of checking: GitHub, OpenAI, Anthropic, Discord, Google Cloud, Azure, Atlassian, Slack, Zoom (routine maintenance only), Vercel, DigitalOcean, Groq, Cohere, Stability AI. The AWS Health Dashboard did not load in our check, so we could not confirm AWS status. Heroku, Stripe, ElevenLabs and Meta pages did not return readable status data, and Cursor and Replicate status pages redirected elsewhere, so those were not confirmed either. Web searches and crowd-sourced trackers showed no corroborating reports of a broad outage or notable BGP incident today.
Overall: incidents detected, limited in scope, with no evidence of a global disruption.