Internet Status — July 21, 2026: No Widespread Outages; AWS Middle East Regions Still Impaired
No internet-wide disruption is underway right now. Checks across crowd-sourced outage trackers and the status pages of the major cloud, CDN, DNS and platform providers show core infrastructure operating normally as of the morning of July 21, 2026.
All clear: Google Cloud, Microsoft Azure, GitHub, OpenAI, Discord, Slack, Atlassian, Vercel, DigitalOcean and Heroku all report full operational status with no active incidents.
Ongoing — AWS Middle East. The AWS ME-CENTRAL-1 (UAE) and ME-SOUTH-1 (Bahrain) regions remain unavailable. These are not new failures: both regions were physically damaged in early March 2026 when data centre facilities were struck during regional conflict, and AWS has said restoration would take several months. Both events are still listed as open on the AWS Health Dashboard, and AWS continues to advise customers to run disaster-recovery plans and shift workloads to alternate regions. Customers outside those two regions are unaffected. (health.aws.amazon.com)
Minor and localised issues. Cloudflare has routine scheduled datacentre maintenance in progress in Baghdad and Dubai, which can cause brief latency increases as traffic reroutes, and continues to investigate degraded availability for two Workers AI models — an issue open since June 16 that does not affect CDN, DNS or WAF delivery (cloudflarestatus.com). Akamai is still monitoring an Application Load Balancer reporting issue that it describes as stable, with all delivery and security services operational. Twilio reports carrier-level SMS and voice delivery problems in Mexico, Malawi and India; these originate in third-party mobile networks rather than Twilio's own platform.
Resolved in the past 72 hours. Meta suffered a broad but short outage on July 19, with Facebook, Instagram and Messenger returning account-unavailable errors for users across North America, Europe and India starting around 03:25 ET; service was largely restored within one to two hours. GitHub resolved a disruption affecting one Copilot model on July 20. DigitalOcean cleared block-storage attachment problems across several regions and an outbound connectivity issue in Bangalore on July 19, and Akamai resolved an Edge Delivery issue affecting India the same day.
One caveat on completeness: Downdetector and Cloudflare Radar both refused automated requests during this run, so the crowd-sourced picture is drawn from secondary trackers and news reporting rather than primary dashboards. Nothing in those secondary sources indicates a significant unreported event.
Bottom line: the internet is up. There is no active multi-provider or internet-wide outage. The AWS Middle East regional outage remains the one substantial open item, and it is a known, geographically contained, long-running recovery rather than a developing incident.