Featured Blog post
As a rule of thumb, the number of on-premise slaves should be calculated in a way that (close to a) hundred percent utilization is accomplished at all times. We’ll configure them so that new slaves are created when needed, and unused ones are destroyed after a short period of inactivity. In this setting, you’ll be able to meet any demand, thus accomplishing high availability and, at the same time, you’ll pay only when slaves are used. In other words, we’ll combine the best of both worlds (on-premise with AWS). Bear in mind that I refer to high availability only in the context of Jenkins slaves. High availability of Jenkins masters is a different problem that can be solved through the High Availability Plugin included in the CloudBees Jenkins Platform.
Let’s get our hands dirty.