Engineering Dashboard

Cloud Cost & Engineering Unit Economics

Streamline AI • Real-time visibility into cost per feature, idle waste, and unit economics

104% Budget
Cost per User
$5.94
+12.7%
MoM Variance
+8.2%
Above Target
Idle Waste
$47K
Critical
Cost per Deploy
$12.40
-3.4%
Tagging Rate
86%
+5%
Budget Usage
104%
Over Budget

Cost per Active User Trend

Target: $5.20

Top Cost Anomaly Services

Budget Consumption

Cloud Spend by Product Line

inference-worker
$142K
34% over budget
api-gateway
$68K
98% of budget
data-pipeline
$53K
web-frontend
$41K
ml-training
$38K
storage
$29K
dev-cluster
$19K
Idle waste

MAU Growth vs Infrastructure Cost

Service Cost Drivers — Last 6 Months

Jan
Feb
Mar
Apr
May
Jun
inference-worker
$98K
$104K
$110K
$126K
$138K
$142K
api-gateway
$61K
$63K
$64K
$66K
$67K
$68K
data-pipeline
$48K
$49K
$50K
$51K
$52K
$53K
ml-training
$32K
$34K
$35K
$36K
$37K
$38K
dev-cluster
$18K
$18K
$19K
$19K
$19K
$19K

Red: >110% budget | Yellow: 90-110% | Green: <90% | dev-cluster shows persistent idle waste

FinOps Recommendations

Recommendation Service Savings Effort Priority
Right-size inference instances (c6i.4xl → c6i.2xl) inference-worker $42K/mo Low P0
Terminate idle dev-cluster nights & weekends dev-cluster $19K/mo Low P0
Migrate to reserved instances (41% → 75% coverage) api-gateway $14K/mo Medium P1
S3 lifecycle: move logs to Glacier after 90d storage $8K/mo Low P1
Enable auto-scaling for data-pipeline workers data-pipeline $7K/mo Medium P1
Spot instances for ml-training (80% workloads) ml-training $11K/mo Medium P2
CloudFront caching improvements (cache hit 62% → 85%) web-frontend $5K/mo Low P2
Tag compliance enforcement (86% → 100%) all-services — Low P2
Total Monthly Savings Potential: $106K
P0 quick wins alone unlock $61K/mo — reducing cost per user to $5.12 (target: $5.20)