Akamai Load Balancer Issue
No widespread internet outage was detected today. Several provider-specific incidents are active, none of them global.
Akamai. Akamai's status page lists an unresolved incident, "Application Load Balancer reporting availability issues" (status: Identified; Akamai says it has found the cause and is deploying a fix). Akamai also published a customer security advisory about an Apache HTTP Server HTTP/2 module vulnerability (CVE-2026-119618) and says it has seen no evidence of exploitation on its platforms. Source: akamaistatus.com
Twilio. Twilio lists multiple regional carrier-level issues: Silent Network Authentication timeouts in the Philippines, voice call failures from some Philippines numbers, one-way audio on some calls from Colombia, SMS delivery receipt delays to New Zealand and Chile, and recovering SMS delays in Brazil and Costa Rica. Planned SoftBank (Japan) maintenance may also delay SMS. These are limited to specific countries and carriers. Source: status.twilio.com
Anthropic (Claude). A partial outage on claude.ai: some users cannot sign up with Apple on the iOS app (mainly new accounts). Google or email sign-up still works. Started Nov 14, still under investigation. Source: status.claude.com
Cloudflare. The status page shows degraded performance for Cloudflare Sites and Services and partial-outage labels on a number of individual data centers, with others under maintenance. No incident text was visible to us and no customer-impact reports turned up in searches, so we cannot say whether end users are affected. Source: cloudflarestatus.com
All clear: AWS, Google Cloud, Azure, Microsoft 365 (business and consumer feeds), GitHub, OpenAI, Discord, Atlassian, Slack, Zoom (only planned maintenance, no impact expected), Vercel, DigitalOcean, Groq, Cohere and Stability AI report no active incidents. Recently resolved: a brief Vercel Sandbox/Blob latency issue (Nov 10) and a DigitalOcean TOR1 storage maintenance overrun (completed Nov 14).
Not verified: Stripe, Heroku, ElevenLabs and Meta status pages did not return readable content, Replicate's page redirected, and Cursor's status address redirected to an unrelated domain, so we did not use it. Our web searches found no reports of a current outage or BGP incident. Downdetector could not be accessed.
Overall: the internet is working normally, with minor provider-specific incidents.