I'm wondering, how do you handle failovers? Is it automatic or manual? On infrastructure I handled from top to bottom, I used VIPs with keepalived (only the v

I'm wondering, how do you handle failovers? Is it automatic or manual? On infrastructure I handled from top to bottom, I used VIPs with keepalived (only the vrrp part, with a weight linked to success/failure of a check script). But in AWS, I'm wondering how to do it properly, maybe DNS records with low TTL (like 1 second).