Skip to main content

Version 1.18.3.5

Release Date: October 1, 2026 Release Type: Stable Previously published as: 1.18.38 (image tag staging-1.18.38)

Fixes for short Redis connection interruptions, alert e-mails on installations with a database read replica, and outage history cleanup on Elasticsearch installations.

Backend Server​

Bug Fixes​

  • Short Redis connection interruptions no longer cause errors: A network interruption of a few seconds between the platform and Redis was reported as a connection error right away, although Redis itself was healthy. Connection and read timeouts are now more tolerant, idle connections are checked so that a dead one is not handed out, and read operations are retried automatically, so a short interruption passes without an error.
  • Alert e-mails are sent on installations with a database read replica: On installations that run the database with a read replica, alert e-mails were never sent, because the step that marks alerts as sent was directed at the read-only replica and rejected there. That step now always runs on the primary database.

Frontend Server​

Bug Fixes​

  • Component outage history is cleaned up on Elasticsearch installations: On installations that use Elasticsearch for monitoring, the database procedures that clean up the component outage history by retention period were never installed. The history was therefore not cleaned up, and Health Check logged a "procedure does not exist" error every minute. The procedures are now installed on these installations as well. Where automatic schema creation is switched off (DATABASE_GENERATE_DDL=false), they have to be created once by hand.

1.18.3.5 is a stable release. Previous: 1.18.3.4