Purchasing strategy under an SLA

advanced

Scenario & brief

A production tier peaks at 6,000 rps and must hold 99.95% availability with critical durability and p99 ≤ 140 ms — on a tight $1,100/mo budget.

Spot looks temptingly cheap, but pure Spot capacity can be reclaimed in bulk, so it can't underpin a 99.95% target on its own — the availability check will fail. The winning move is a Savings Plan (or On-Demand) for the steady baseline, the right Graviton class on an autoscaling group across two AZs, an ElastiCache tier to keep the database in budget, and Multi-AZ RDS with backups.

6,000 rps peakp99 ≤ 140ms99.95% availdurability: criticalbudget $1,100/mo

App Servers

Compute

Instance class

Scaling

Instances (static)

6instances

Availability Zones

Purchasing

System health

Erupting · SLA breach

/ 100

Score

SLA not met yet

Monthly cost

$2,054

Budget $1,100/mo · $954 over

Metrics

Capacity57

Availability30

Durability25

Cost efficiency8

Requirements

Peak capacity — 21600 rps compute · 3400 rps db (need ≥ 6000 rps)
p99 latency — ~132 ms (need ≤ 140 ms)
Availability — 99.00% (need ≥ 99.95%)
Durability — at risk (need redundancy + backups)
Budget — $2054/mo (need ≤ $1100/mo)

Advisor

The database is saturated at peak — add a cache to shed read load, or scale it up.
Compute runs in a single AZ — spread across ≥2 AZs (with ≥2 instances) to meet the availability target.
The database has no Multi-AZ standby or replica — a failure risks data loss. Enable Multi-AZ and backups.

Discussion

No comments yet. Be the first to start the discussion.

For learning purposes only. Costs and capacities are illustrative, not live AWS prices. Not affiliated with or endorsed by Amazon Web Services.