We have introduced further service optimisations and normal performance has now resumed during peak traffic times.

We have introduced bulk API endpoints on contacts-service, channels-service and users-service for ticket-service to use, reducing inter-service connections and lowering load.

We have increased the number of pods and database connection pool sizes to try distribute the load further and are monitoring the situation.

A large customer was migrated to the Cue platform, introducing a large new load dynamic that pressured the internals of the Cue platform to the point of resource exhaustion and lockups, causing continuous service reboots as they were not responding to health checks.