Daily digest · 2026-10-31 (UTC)

What broke for AI agents on 31 Oct 2026

1 incident on 1 status page.

Final. Built 2026-11-01 07:50 UTC, 7 hours after the day ended; this page will not change.

What broke

Cloudflare: “Cache purge delays in Qingdao Shandong, China”minor impact · opened 06:41 UTC · resolved 06:57 UTC (16 minutes) · CDN Cache Purgestatus page ↗provider-incidents v82

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

9 other releases on the day

@anthropic-ai/claude-agent-sdk 0.3.316 (npm)released 00:39 UTC · previous 0.3.314agent-sdk-releases v96
anthropics/claude-agent-sdk-python v0.2.176published 01:47 UTC · previous v0.2.175release notes ↗agent-tool-releases v96
anthropics/claude-code v2.1.316published 01:01 UTC · previous v2.1.314release notes ↗agent-tool-releases v96
claude-agent-sdk 0.2.176 (pypi)released 01:45 UTC · previous 0.2.175agent-sdk-releases v96
openai 3.42.0 (pypi)released 00:27 UTC · previous 3.41.0agent-sdk-releases v96
openai 7.47.0 (npm)released 05:06 UTC · previous 7.45.1agent-sdk-releases v96
openai/openai-node v7.46.0published 00:27 UTC · previous v7.45.1release notes ↗agent-tool-releases v96
openai/openai-node v7.47.0published 04:46 UTC · previous v7.46.0release notes ↗agent-tool-releases v96
openai/openai-python v3.42.0published 00:26 UTC · previous v3.41.0release notes ↗agent-tool-releases v96

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 v128
477 of 500 remote MCP endpoints probed answered a handshake (95.4%), on 334 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 v30

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.