Zhivko Todorov
ALL CASE STUDIES

CASE 46 · QUARTZ · 2024

REDSHIFTRA3RESERVEDCONCURRENCY SCALING

Redshift, sized for what the team actually queries.

An analytics platform had been on Redshift DC2 nodes since 2019. The cluster was provisioned for the peak Monday-morning load and ran at 11% utilisation the rest of the week. We migrated to RA3 with managed storage, added Reserved capacity at the right tier, and let Concurrency Scaling absorb the peaks.

INDUSTRY

Analytics platform

DOMAIN

COST

DELIVERED

2024

STACK

REDSHIFT RA3·CONCURRENCY SCALING·REDSHIFT SPECTRUM·S3·RESERVED CAPACITY·CLOUDWATCH METRICS

RESULTS

What changed, by the numbers.

CLUSTER COST

−55%

$34K → $15K / MONTH

STORAGE

UNCOUPLED

NOW PAY-AS-USED

PEAK QUERY LATENCY

−18%

CS BURST

OFF-HOURS USAGE

PAUSED

SCHEDULED

HOW IT WENT

The DC2 cluster had been sized in 2019 by an engineer who no longer worked there. Data volume had grown; query patterns had changed; nobody had revisited. Monday at 9am was painful; Sunday at midnight was a $5,000 idle cost.

RA3 separated compute from storage and let us right-size for the median, not the peak. Concurrency Scaling absorbed Monday-morning load (and got us through quarterly close week). Reserved capacity covered the always-on compute. Off-hours pause was scheduled via EventBridge.

Bill dropped 55% by month two. Peak query latency actually improved 18% — Concurrency Scaling cut the queue depth that DC2 had been silently building. The off-hours pause saves another $4k/month without any user impact (the off-hours users are scheduled jobs that move to next-morning).

READY WHEN YOU ARE

Let's get your AWS bill (and architecture) in order.

The discovery call is free. You walk away with at least one concrete idea — even if we never work together.

Or email directly →