Topic
Databases
Managed databases trade operational effort for premium pricing, which makes provisioning discipline the whole game.
AWS's managed database services remove genuinely hard operational work: patching, failover, backups, replication. The premium you pay for that is worth it for most teams. What the premium does not forgive is provisioning against imagined load instead of measured load.
Databases hold their waste in different places than compute. Instances sized for launch-day estimates that were never revisited. Multi-AZ deployments on development databases. Provisioned IOPS nobody consumes. Aurora capacity floors set well above the workload's actual trough. And because databases feel risky to touch, these decisions calcify.
The material here covers engine and pricing-model choice, reading utilization honestly, and the safe paths to resizing the databases everyone is afraid to touch.
Nothing here yet
Research on databases lands here soon. Browse the rest of the library in the meantime.
Related topics
Compute
EC2, Lambda, and Fargate. Instance families, rightsizing, and the price of idle capacity.
Commitments
Reserved Instances and Savings Plans. Coverage, utilization, and sizing commitments to real usage.
Storage
S3, EBS, and archival storage. The classes, lifecycle rules, and hidden retrieval costs that decide the bill.