Cluster testing evaluates whether a database cluster can withstand production-level traffic and meet operational requirements. Treat the result as a baseline to tune rather than a perfect initial design. section 9.2.2
Use testing to guide multidimensional cluster decisions:
- Send consequence-free, production-level traffic and observe whether the cluster withstands the workload. section 9.2.2
- Account for disk space, CPU cores, memory, network bandwidth, and cost rather than optimizing one resource in isolation. section 9.2.2
- For an estimated 1 TB of data with replication factor 3, allocate 6 TB across the cluster to allow for replication, uncompacted data, and growth. section 9.2.2
Figure 9.5 Using smaller nodes in your cluster can be a cost-effective way to meet your throughput goals, as well as increase your cluster’s availability if you were to lose a quorum.
Compare smaller-node layouts when testing throughput goals and resilience after losing a quorum. section 9.2.2
