Drop problematic background jobs from the Sidekiq queue
This uses the Sidekiq API to drop `PatchPrometheusServicesForSharedClusterApplications` because this job can blow up Sidekiq queues by retrying excessively. Background jobs are intended to process many rows for a large batch of records, but this migration tends to process a single row within a second. As a result, the exclusive lease causes this job to rescheduled over and over again. Identified in https://gitlab.com/gitlab-com/gl-infra/infrastructure/issues/9064
Showing
Please register or sign in to comment