Maintenance Window

* Scheduling Guide

Maintenance window schedules coordinate planned downtime and housekeeping tasks during the lowest-traffic period of the week. Database vacuum and reindex operations are the most common examples — PostgreSQL's VACUUM ANALYZE and MySQL's OPTIMIZE TABLE are typically run nightly or weekly depending on table churn rate. The window should be long enough to accommodate the slowest expected operation with headroom, and jobs should exit gracefully if they are still running when the window closes. Kubernetes CronJob is a common choice for maintenance tasks because it enforces a deadline via activeDeadlineSeconds and can be scheduled with concurrency policies that prevent overlap. Communicate the maintenance window to downstream systems so dependent jobs do not queue up requests during the downtime.