Capacity planning is the practice of estimating future resource needs from current usage, growth trends and upcoming events, then provisioning enough headroom before demand arrives. It covers CPU, memory, disk, network, database connections and third-party limits.
Why it matters
Running out of capacity looks like an outage: slow pages, timeouts and 503 errors under load. Adding capacity during an incident is slow and risky, while adding it ahead of time is routine. Planning also prevents the opposite waste of paying for large servers that sit idle.
Autoscaling helps but does not replace planning. It reacts after load rises, has startup delay, and cannot exceed quotas, database limits or the capacity of downstream services.
Example
A service peaks at 400 requests per second on 4 servers running at 55% CPU. Load testing shows each server stays healthy up to about 70% CPU. Traffic is growing 8% per month and a campaign may double peak load.
current peak CPU capacity used: 4 x 55% = 2.2 servers
campaign doubling: 4.4 servers
target 70% utilization: 4.4 / 0.70 = about 6.3 -> provision 7The numbers are illustrative. Use your own measured limits.
Common mistakes
- Planning for average load instead of peaks and events such as sales or launches.
- Forgetting the weakest link, often the database or a payment API.
- Skipping load tests, so the headroom figure is a guess.
Related
Get notified when this term updates
Plus occasional practical tips on uptime and monitoring, nothing else. Unsubscribe anytime.
This page is general information written by our editorial team and updated when documentation changes. Servers and providers differ, so test changes safely first. See our editorial policy and disclaimer.