Daily digest · 2026-10-28 (UTC)

What broke for AI agents on 28 Oct 2026

7 incidents on 3 status pages.

Final. Built 2026-10-31 11:03 UTC, 7 hours after the day ended; this page will not change.

What broke

1. GitHub: “Disruption with some Copilot services”critical impact · opened 14:12 UTC · resolved 15:28 UTC (75 minutes) · Copilotstatus page ↗provider-incidents v73
2. Anthropic: “Elevated errors affecting the Gmail connector”minor impact · opened 03:59 UTC · resolved 05:48 UTC (109 minutes) · Claude Code | Claude Cowork | claude.aistatus page ↗provider-incidents v71
3. Anthropic: “Elevated errors affecting the Google Calendar connector”minor impact · opened 16:19 UTC · resolved 20:05 UTC (3 h 46 min) · Claude Code | Claude Cowork | claude.aistatus page ↗provider-incidents v74
4. Cloudflare: “Increased Errors for Durable Objects in the Eastern North America region”minor impact · opened 00:17 UTC · resolved 03:28 UTC (3 h 12 min) · D1 | Durable Objects | Queues | Workflowsstatus page ↗provider-incidents v71
5. Cloudflare: “Increased Errors for Durable Objects in the Eastern North America region”minor impact · opened 09:45 UTC · resolved 11:10 UTC (86 minutes) · D1 | Durable Objects | Queues | Workflowsstatus page ↗provider-incidents v72
6. Cloudflare: “Increased latency and error rate for R2 in ENAM”minor impact · opened 21:01 UTC · resolved 21:12 UTC (11 minutes) · Artifacts | Durable Objects | R2status page ↗provider-incidents v74
7. Cloudflare: “Network Performance Issues in Ashburn, IAD”minor impact · opened 04:39 UTC · resolved 22:18 UTC (17 h 39 min) · Ashburn, United States - (IAD)status page ↗provider-incidents v74

Shipped

No new major version of the SDKs and tools we follow.

19 other releases on the day

1. @langchain/core 1.3.1 (npm)released 00:56 UTC · previous 1.3.0agent-sdk-releases v85
2. @mistralai/mistralai 3.4.0 (npm)released 15:25 UTC · previous 3.3.0agent-sdk-releases v87
3. ai 7.0.156 (npm)released 18:35 UTC · previous 7.0.155agent-sdk-releases v88
4. BerriAI/litellm v1.105.7published 22:55 UTC · previous v1.106.5release notes ↗agent-tool-releases v88
5. BerriAI/litellm v1.106.6published 23:03 UTC · previous v1.105.7release notes ↗agent-tool-releases v88
6. langchain-ai/langchain langchain-core==1.7.1published 15:00 UTC · previous langchain-core==1.7.0release notes ↗agent-tool-releases v87
7. langchain-core 1.7.1 (pypi)released 14:59 UTC · previous 1.7.0agent-sdk-releases v87
8. litellm 1.106.6 (pypi)released 22:48 UTC · previous 1.106.5agent-sdk-releases v88
9. mistralai 3.6.0 (pypi)released 09:23 UTC · previous 3.5.0agent-sdk-releases v86
10. modelcontextprotocol/servers v2.2.0published 20:57 UTC · previous v2.1.0release notes ↗agent-tool-releases v88
11. ollama/ollama v0.40.13published 19:55 UTC · previous v0.40.12release notes ↗agent-tool-releases v89
12. openai 3.41.0 (pypi)released 22:27 UTC · previous 3.40.0agent-sdk-releases v88
13. openai/openai-python v3.41.0published 22:26 UTC · previous v3.40.0release notes ↗agent-tool-releases v88
14. vercel/ai @ai-sdk/[email protected]published 18:36 UTC · previous @ai-sdk/[email protected]release notes ↗agent-tool-releases v88
15. vercel/ai @ai-sdk/[email protected]published 18:36 UTC · previous @ai-sdk/[email protected]release notes ↗agent-tool-releases v88
16. vercel/ai @ai-sdk/[email protected]published 18:36 UTC · previous @ai-sdk/[email protected]release notes ↗agent-tool-releases v88
17. vercel/ai @ai-sdk/[email protected]published 18:36 UTC · previous @ai-sdk/[email protected]release notes ↗agent-tool-releases v88
18. vercel/ai [email protected]published 18:34 UTC · previous [email protected]release notes ↗agent-tool-releases v88
19. x402 2.26.0 (pypi)released 18:29 UTC · previous 2.25.0agent-sdk-releases v88

Measured on the day

1. 184 round-trips to 23 endpoints agents depend on, 0 failedmodel APIs, registries, chains, the x402 facilitator, status pages and two baselines · measured from macbook-kragent-api-observatory v116
2. 458 of 500 remote MCP endpoints probed answered a handshake (91.6%), on 471 hostsa different slice of the official registry each day, so the share is not compared across days · one attempt per endpoint, an auth challenge counts as an answer · measured from macbook-krmcp-server-liveness v27

How this is made

• Every line comes from one of WITAN's public datasets and links the version its records landed in: provider-incidents, agent-api-observatory, mcp-server-liveness, agent-sdk-releases, agent-tool-releases. They are free to read.
• Incidents are what the providers' own status pages published (Anthropic, OpenAI, GitHub, Cloudflare, npm, Vercel), read every 6 hours; scheduled maintenance is left out. Follow a line's status page link for the provider's text.
• Round-trips and MCP handshakes are WITAN's own measurements from one vantage point, named on each line. A failure there can be the network between us and the service; it is not a verdict on the service. A latency line needs the day's median at least 2× the previous 7 days' and 250 ms more.
• Releases are those of the SDKs and agent tools the datasets follow; pre-releases are left out.
• Days are UTC. A day is rebuilt every hour while it runs and frozen 7 hours after it ends. Model prices are not here: that dataset's collector is off. WITAN is a testnet preview run by one developer.

Earlier days

2026-11-012026-10-312026-10-302026-10-292026-10-282026-10-272026-10-262026-10-252026-10-24