Capacity Planning in Four Parts: Telling the Future without a Crystal Ball
summary
Four clear steps: pick an SLO and translate business risk into an error budget, find driver metrics by aligning SLIs with resource inflection points, turn those into actionable thresholds and dashboards with non‑paging alerts and collapsible detail for vetting, and project forecasts that include lead time and buffers. Case study shows a reactive pattern of 16 servers every 64k users leaving many idle and suggests moving toward 1 per 4k units plus a 10% buffer. Warns against naive driver formulas and heavy daily linear regressions.
AI-written summary · about
chapters (6)
- 0:00Start
- 1:35CAPACITY PLANNING OBJECTIVE
- 1:53A CASE STUDY
- 3:06The Hard
- 4:33CASE STUDY: The Ugly
- 17:58SOME FINAL TIPS
related talks
27mHow to lower the costs of your Drupal Site's resources and plan...
06mPrevent, Rather than Fix: Scaling to Enable Customers
09mHow to answer tricky questions the SRE way
30mTechniques for SLOs and Error Budgets at Scale
41mSLOs that Delight: Master Multi-Service SLO Calculations
43mSLO Burn—Reducing Alert Fatigue and Maintenance Cost in Systems of Any Size
29mNot All Minutes Are Equal: The Secret behind SLO Adoption Failure
38mA Practical Look at Performance Theory