Docs
    AWS can schedule events for your EC2 and RDS instances. These events cannot be canceled and are scheduled for various maintenance related reasons. AWS sends notifications of upcoming scheduled events before they take place.


    EC2 Scheduled Events


    EC2 scheduled events provide a description of the event and a window that the event can occur within. EC2 can schedule the following event types:

    • Instance stop: the instance will be stopped and will be migrated to a new physical host when you restart it.
    • Instance retirement: the instance will need to be replaced.
    • Reboot: either the instance will be rebooted or the underlying physical host will be rebooted.
    • System maintenance: the instance might be temporarily affected during maintenance to the datacenter

     

    In all cases, your EC2 instance will be unavailable while the scheduled event is occurring. To immediately resolve a scheduled event, you can start an additional instance to replace the affected one, and then terminate the affected one. This should be done at a time when it will the least effect on your service. Or, if your app does not require high availability, you can just take action on the event whenever you deem best.

     

    RDS Scheduled Events


    RDS scheduled events are scheduled for updates to the OS your database instance is running on, or for updates to the database engine of your instance. These updates can be applied immediately, deferred, or scheduled to be run during your next maintenance window. Some updates can be deferred indefinitely, while others are required.

     

    To better handle RDS scheduled events, run your instances in Multi-AZ mode. If the required updates are OS updates, they will be applied to the standby instance first, after which the standby will handle traffic while the primary is being updated. Updates to the database engine are applied to both instances at once and the database will be unavailable for a brief period of time.

     

    Resources