OpenAI Latency, Spotify Recovered
No internet-wide outage is in progress as of about 20:20 UTC on September 30, 2026. AWS, Google Cloud, Azure, Azure DevOps, GitHub, Anthropic, Discord, Slack, Atlassian, Vercel, Heroku, DigitalOcean, Stripe and the smaller AI platforms we track all report normal operation. Today's open incidents are small and mostly recovering.
**OpenAI: elevated API latency (recovering)**
OpenAI opened an incident at 17:05 UTC for elevated latency on some API requests, with Chat Completions and Responses marked degraded. A mitigation is in place and OpenAI is monitoring recovery. A separate issue with elevated errors in ChatGPT Space Pages, open since 02:20 UTC, is also in monitoring. Source: https://status.openai.com
**Cloudflare: Madrid performance and Workers build delays (monitoring)**
Cloudflare reported network performance issues in Madrid from 10:31 UTC and delays to Cloudflare Workers builds from 15:33 UTC. Fixes are deployed for both and Cloudflare is monitoring. The Asia-Pacific network degradation that Cloudflare had open since September 23 no longer appears among its active incidents. Two older low-impact issues remain open: Cloudflare One clients being wrongly challenged on some sites (since September 22) and WARP geolocation errors (since August 27). Source: https://www.cloudflarestatus.com
**Akamai: APAC origin errors linger**
Akamai's Canada edge delivery incident from September 28 is no longer listed. Its APAC origin connectivity incident stays in monitoring while an upstream carrier finishes repairs on the damaged subsea cables. Akamai says immediate impact has subsided, but intermittent errors are possible until the work is done. Source: https://www.akamaistatus.com
**Twilio and Zoom: carrier-level issues**
Twilio is tracking SMS failures from some short codes to several networks in Chile (recovering) and SMS delays and failures to Metfone, Viettel, Beeline and Sotelco subscribers in Cambodia (cause identified). Zoom is still working with a provider on Polycom desk phones that went offline on September 25, which is affecting Zoom Phone devices. Sources: https://status.twilio.com, https://www.zoomstatus.com
**Recently resolved: Spotify outage (September 29)**
Spotify went down for a large number of users on Tuesday, September 29. Downdetector reports started climbing around 9:30 to 9:45 AM ET (13:30 to 13:45 UTC) and peaked near 10:41 AM ET. Downdetector received more than 200,000 reports from 173 countries, about 46% of them from the US. Most complaints were about the mobile app (roughly 58%) and audio streaming (21%). Service came back later that morning. Spotify has not given a public cause. This happened after our September 29 report was filed, so it appears here. Source: https://www.forbes.com/sites/zacharyfolk/2026/09/29/spotify-outage-thousands-of-users-report-problems-with-music-streaming-app/
**Other notes**
Crowd-sourced trackers show scattered reports for LinkedIn and Grok today, but neither company has confirmed a problem and we could not verify real impact. Microsoft retires the Azure NVv4 GPU VM series today. That is a scheduled retirement, not an outage. Meta's status page could not be read this run, and we found no reports of a Meta disruption.
**Bottom line:** the internet is up. Today's incidents are regional or limited to single products, and most are already recovering.