Control D DNS outage: paid plans stopped resolving for about 54 minutes
Control D's DNS service for paying customers stopped answering queries at every location on October 11, 2026, according to an incident report the company has published. The outage reportedly ran from 16:40 to 17:35 UTC, roughly 12:40 to 1:35 p.m. EDT. For affected households it may have looked like the internet itself had gone down.
What's happening
From about 16:42 UTC, DNS queries from paying customers went unanswered everywhere, and service began returning at 17:31 before full restoration at 17:35. The vendor says its software returned no answer rather than a wrong one, so queries were not allowed, blocked or redirected against customers' filtering settings.
Control D attributes the failure to a format mismatch in its own API. An update released on September 16 added validation for an optional per-domain override in "services", the maintained domain lists behind entries such as streaming platforms. As part of that change, the API reportedly began storing the override as an integer, while the DNS software expects a string.
That code path only ran when a domain was saved with an explicit override such as "allow" or "block". The first came on October 11 at 16:40 UTC, when a team member added four domains to one service with an "allow" override. Because a service's domain list is shared configuration, every DNS server for paid plans reloaded at once. None could read the new values, and an error-handling flaw made the software stop instead of logging and skipping the unreadable entry.
A safeguard that shifts traffic to healthy locations could not help: every location had the same problem. The software was also designed not to start when it could not read all of its configuration, so automatic restarts failed on the same values until they were corrected.
Detection was slow: the first automated alerts at 16:46 named individual servers rather than a full outage for paying customers, and on-call was paged at 16:51. Control D posted publicly on Reddit and Discord between 17:05 and 17:16, and traced the cause at 17:16. The four overrides were removed at 17:30 and service was fully restored by 17:35. Control D had no incident status page at the time.
Who's affected
The incident affected DNS resolution for all paid plans, listed as Personal and Organization accounts, at every location and on every protocol, including DNS-over-HTTPS and DNS-over-TLS. It covered every app, router, device and integration on them. The vendor's free DNS service reportedly stayed online.
Households where every device points at Control D may have appeared to lose internet access, including phones on mobile data with a persistent profile. Organisations that use the service as their primary resolver may have lost name resolution across their endpoints.
Control D states that no query was answered incorrectly and that customer settings and data were not changed or exposed. It has not said how many customers were affected, so the exact scale remains unknown. The outage occurred on October 11, 2026, while the report circulated publicly around October 20, so some readers may only now hear about it.
What to do
- If your devices failed to resolve names during that window and recovered afterwards, there is likely nothing to fix locally: the vendor calls the fault server-side and says it is corrected.
- Review your profiles, filters and custom rules. Control D says settings were not changed, but a quick check of the dashboard is a reasonable precaution after a provider incident.
- If you rely on a single DNS provider, consider a secondary resolver or fallback path. The vendor notes its free service stayed online but has not published a formal workaround for paid plans.
- If problems appear to continue, note the time, location, protocol and the device or router involved, then report it to Control D support.
- Watch the vendor's own channels. A hosted status page is being set up; until then, its blog and community channels appear to be the primary sources.
The immediate bug is reportedly fixed: the DNS software now sets aside and logs unreadable entries instead of stopping, and the API stores overrides as text again. Further work listed by the company includes configuration validation, staged rollouts and extra alerting.
This is an evolving situation and details may change. The account rests on Control D's own report; independent confirmation of its internal timeline is not available.
Sources
- Control D — Incident Report: DNS Outage on October 11, 2026 — The vendor's primary account of the outage window, root cause, timeline and corrective actions.
- Hacker News — discussion of the incident report — The submission that surfaced the report publicly on October 20, 2026.