This topic describes optimizing a workload with Headroom reduction.
A workload is optimized after the following steps:
You apply Headroom reduction to a workload and choose whether Kompass should automatically reduce and manage minimum replicas or whether you will do that manually.
Kompass prepares and hibernates nodes to protect the workload.
For auto-managed optimization, Kompass reduces the workload minimum replicas.
This process completes within an hour of applying the solution.
From the user interface, you can choose from the following minimum replica optimization strategies:
Strategy | Workload protection | Minimum replicas | Savings benefit | Relevant for this kind of workload |
Auto-managed | Full | Reduced to the lowest safe level | Highest savings | Most workloads |
Self-managed | Full | Managed manually by the user (not reduced by Kompass) | Not applicable | Testing or custom-configured workloads |
There is an additional automated optimization strategy - Conservative - that can be applied using a CRD, not from the user interface. The conservative strategy provides more moderate reductions in minimum replicas and savings, and is more relevant to sensitive or stateful workloads.
After Kompass reduces the minimum replicas in either of the automated strategies, Kompass continues to analyze and adjust - if needed - the minimum replica setting to ensure that managed workloads are being optimized according to the selected strategy. Changes are made when a change in the usage pattern is determined or when an optimization strategy is changed.
To apply Headroom reduction to a workload, see Apply Headroom reduction to a workload.
After Headroom reduction is active on a workload, you can see the impact on the workload in terms of CPU, RAM, cost, and other metrics. For more info, see Apply and manage Headroom reduction.
For information about supported environments and limitations, see Compute solution supported environments and limitations.
See also: