Predictive auto-scaling
If you manually dial up and down the node counts and auto-scale on schedule or metric triggers, you still will have some issues bringing up services you need at exactly the time you want them to run since services take a bit of time to get online, self-configure, and start getting propagated to various load balancers in your network. With that type of architecture, it is likely that your users will be the one discovering that you do not have enough capacity and your system then reacting to compensate. If you are really striving for all-out best user experience from your services sometimes you may also need to add one more layer to your auto-scaling triggers that can predict when your service will need more resources ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access