The internet is up globally as of 10:00 UTC on Monday, October 5. There is no widespread outage. What's active this morning is limited to specific products, and the most notable item from the last 24 hours is a regional AWS incident in Spain that has since been resolved.
**Ongoing**
**OpenAI (ChatGPT Work Mode):** OpenAI opened an incident for elevated Work Mode errors at 07:03 UTC. ChatGPT Work and Codex in the ChatGPT desktop app show degraded performance, and the incident is in monitoring, which means a fix has been applied. Core ChatGPT and the API are not listed as affected. It follows a busy week at OpenAI, including a major login and signup incident on October 1, now resolved. [status.openai.com](status.openai.com)**Twilio (Europe messaging):** Twilio opened two new messaging incidents overnight: delayed messaging logs and status callbacks in its IE1 (Ireland) region, now in monitoring, and SMS delivery delays and failures between Twilio long codes and several French mobile networks, with the cause identified. Older MMS issues in Australia and with GCI in the US remain open. [status.twilio.com](status.twilio.com)**Long-running, low-impact:** Akamai's APAC origin connectivity issue (since September 22) is still in monitoring. Zoom's Polycom device issue (since September 25) is still open. Cloudflare has two minor open items, a Cloudflare One challenge issue and incorrect WARP geolocation, plus routine data-center maintenance and rerouting.**Resolved in the past 72 hours**
**AWS EU-SOUTH-2 (Spain):** Between 15:48 and 18:28 UTC on October 4, a single Availability Zone (eus2-az1) saw elevated packet loss. AWS said this also raised API latencies and error rates for multiple services in the region. AWS found the root cause and the issue is fully resolved. [health.aws.amazon.com](health.aws.amazon.com)**Philippines subsea cable cut:** The October 2 damage to the PhilippinesâSingapore and PhilippinesâHong Kong links slowed international traffic for about two days. PLDT reported full restoration on October 4, and Globe and Converge are carrying traffic on alternate routes. The physical repair may take weeks, and DICT is still investigating the cause. [GMA News](gmanetwork.com) · [Rappler](rappler.com)**Cloudflare:** CDN configuration changes and dashboard cache purges briefly returned errors on October 2 (about 11 minutes; API purges were unaffected), and Workers Builds failed to start on October 3. Both are resolved. [cloudflarestatus.com](cloudflarestatus.com)**Regional and unconfirmed signals**
Outage-tracking sites showed a rise in AT&T user reports from around 04:30 UTC (00:30 ET). It looks like hundreds of reports, not thousands, and we found no confirmation from AT&T, so treat it as unverified.IODA flagged a connectivity drop in Kampong Speu province, Cambodia, at 04:35 UTC. It comes from a single source and is low-confidence. [Voidly/IODA](voidly.ai)Internet Society Pulse lists 10 government-ordered shutdowns ongoing worldwide. [pulse.internetsociety.org](pulse.internetsociety.org)**All clear**
AWS (outside the long-standing Middle East region outages), Google Cloud, Azure, Azure DevOps, GitHub, Anthropic/Claude, Discord, Atlassian, Slack, Vercel, Heroku, DigitalOcean, Stripe, Cursor, Groq, ElevenLabs, Cohere and Stability AI all report no active incidents. Replicate's status page now redirects into Cloudflare's, which lists Replicate as operational. Meta's status page shows no incident we could read.
Status: incidents detected. These are product-level issues, and there is no internet-wide disruption.