Beginning at ~14:40 UTC on December 8, 2025, a subset of Compute hosts began to experience errors resolving DNS names including a core Linode database endpoint. Most Compute hosts had an active database connection and were able to process new jobs without impact. At that time a separate deployment of a new internal service was in process which required restarting the service and the re-establishment of database connections. Due to this restart, these hosts disconnected from the database and were unable to reconnect due to the DNS resolution problem.
This issue was due to a recent infrastructure code configuration update to switch these hosts to internal DNS resolvers over IPv6. The change was intended to be applied only to storage systems and unintentionally was applied to all hosts including Compute hosts.Once the offending configuration change was isolated, it was reverted. To resolve the issue, we worked through the impacted hosts, updating configurations in batches, so as to not introduce additional unnecessary impact, fully mitigating the issue at ~21:17 UTC on December 8, 2025.
To prevent this issue from occurring in the future, we will decouple the currently-shared configurations whose update led to this impact.
This summary provides an overview of our current understanding of the incident given the information available. Our investigation is ongoing and any information herein is subject to change.