☁️ DevOps & Cloud · Platform & Delivery
Plan capacity from growth curves, not incidents
Headroom policies, load-shed tiers defined calmly, and procurement lead times respected.
advanced~40 minDevOps EngineersSREsCloud Architects
Steps
- 1Forecast demand from sales pipeline + seasonality, reviewed monthly
- 2Set headroom policy per tier (e.g., 40% at peak) with alert thresholds below it
- 3Load-test at forecast+headroom levels, not last year's peak
- 4Pre-negotiate procurement/reservation lead times into planning calendars
- 5Define load-shed ladder in advance: which features drop first, who decides
- 6Post-incident capacity reviews update curves and buffers formally
Common Pitfalls
- ▲Capacity added reactively at 95% utilization
- ▲Shed-ladder improvised during the outage it was meant for
Commands
Install with skills CLI
$ npx skills add aniruddhaadak80/skills --skill platform-delivery-capacity-planning-forecastingInstall globally
$ npx skills add aniruddhaadak80/skills --skill platform-delivery-capacity-planning-forecasting -gTags
#capacity#planning#scaling#devops-cloud#platform-delivery