Daily digest · 2026-11-12 (UTC)

What broke for AI agents on 12 Nov 2026

6 incidents on 2 status pages.

In progress. As of 17:48 UTC, rebuilt every hour; it freezes at 07:00 UTC on 13 Nov 2026.

What broke

Cloudflare: “Changes to Cloudflare Workers are not producing Event Subscription events”minor impact · opened Nov 10 01:53 UTC · identified as of 17:48 UTC · Queues | Workersstatus page ↗provider-incidents v111
Cloudflare: “Replicate Model Registry Unavailable”minor impact · opened Nov 11 04:33 UTC · identified as of 17:48 UTC · Replicatestatus page ↗provider-incidents v114
Cloudflare: “DNSSEC Configuration Change Issues”minor impact · opened 04:53 UTC · resolved 05:26 UTC (33 minutes) · DNSSEC Updatesstatus page ↗provider-incidents v115
Cloudflare: “Replicate Issue”minor impact · opened 08:30 UTC · resolved 11:34 UTC (3 h 4 min) · Replicatestatus page ↗provider-incidents v116
OpenAI: “GPT-image-2.5 and 2.6 Sunburst & Flare High error rate”the feed gives no impact level · the feed gives no start time · resolved 09:25 UTC · Imagesstatus page ↗provider-incidents v116
OpenAI: “Unable to start new Workspace Agent runs”the feed gives no impact level · first seen by our reader 00:20 UTC (the feed gives no start time) · resolved 08:32 UTC · Agentstatus page ↗provider-incidents v116

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

4 other releases on the day

BerriAI/litellm v1.106.9published 01:06 UTC · previous v1.106.8release notes ↗agent-tool-releases v131
BerriAI/litellm v1.108.6published 01:12 UTC · previous v1.106.9release notes ↗agent-tool-releases v131
BerriAI/litellm v1.109.1published 02:07 UTC · previous v1.108.6release notes ↗agent-tool-releases v131
litellm 1.109.1 (pypi)released 01:59 UTC · previous 1.109.0agent-sdk-releases v133

138 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 v175
475 of 800 remote MCP endpoints probed answered a handshake (59.4%), on 517 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 v42

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.