Stage 7: Orbital Maneuvering — Respond to Demand
More passengers are searching for flights. Before changing replicas, adding a cache, or tuning resources, we need a baseline: what workload ran, what changed, which result improved, and what else became worse? Orbital Maneuvering treats performance as an investigation rather than a collection of magic knobs.
Caching changes the work a request performs. HPA changes the desired replica count from measured demand. VPA recommends or changes resources. Node capacity and scheduling decide whether new Pods can actually run. These mechanisms often appear together, but they solve different parts of the problem.
What you will understand
As you read, ask what each mechanism measures, which value it changes, and what must already exist for that change to help. The chapters introduce the precise autoscaling calculations and configuration only after the performance question is clear.
When to take the controls
Open the Orbital Maneuvering lab once you can state the baseline and predict whether a cache, extra replicas, or a resource recommendation is the relevant response. Then use the local experiment to collect evidence rather than just watch numbers move.