TL;DR: Buoyant Cloud helps engineering teams identify and eliminate GCP cost waste through architecture-level fixes — not dashboards. Amit Malhotra has recovered $15,000–$25,000/month in addressable waste on a typical GCP audit. US and Canada.

GCP FinOps & Cloud Cost Optimisation — Find and Fix the Waste

Many teams assume rising GCP costs come from growth. In reality, cloud waste often comes from infrastructure decisions that quietly compound over time — oversized GKE clusters, unmanaged non-production environments, unnecessary NAT traffic, and resources nobody owns.

GCP FinOps is not just about visibility into spend. It’s about identifying the architectural patterns creating waste and fixing them before costs scale further.

Why Your GCP Bill Keeps Climbing (And Nobody Can Explain It)

A Series B SaaS team thought their GCP spend was under control—steady at $18,000/month.
Then it climbed to $47,000.
– No major launches.
– No traffic spike.
– No clear reason.

Dashboards showed everything as “normal.” Compute, storage, and usage patterns hadn’t changed in any obvious way.
But the cost kept increasing.

The problem wasn’t visibility—it was architecture.

Internal traffic routing through NAT, over-provisioned GKE node pools running 24/7, and non-production environments using production-sized resources were quietly driving up costs.
Individually, these don’t trigger alerts. Together, they create significant, compounding waste.

This isn’t a one-off issue—it’s a pattern across scaling teams on GCP.

The 6 Places GCP Waste Hides

1. NAT Gateway charging per-GB on internal traffic: Explanation + fix (VPC Endpoints / Private Service Connect)

2. GKE node pools sized for peak load, running 24/7: Explanation + fix (Cluster Autoscaler, node auto-provisioning, right-sizing, GKE Autopilot for variable workloads)

3. Cloud SQL running production-sized instances in Dev/Staging: Explanation + fix (schedule stop/start outside business hours, right-size to db-f1-micro for non-prod)

4. Log exports sending everything to BigQuery: Explanation + fix (log filters, export only what matters, lifecycle management on log buckets)

5. Zombie resources from deprecated projects: Explanation + fix (billing export to BigQuery, 1budget alerts per project, Organisation Policy to enforce labels)

6. No one owns the bill: Explanation + fix (FinOps ownership model, budget alerts by service, weekly cost review cadence)

How Buoyant Cloud Approaches GCP FinOps

Not a dashboard configuration. An architecture review.


The fastest way to reduce GCP cost is to fix the architectural decisions causing waste — over provisioned compute, misrouted network traffic, unmanaged non-production databases.


Buoyant Cloud’s GCP cost audit covers:


→ Billing export analysis — identify the exact line items driving cost
→ GKE right-sizing assessment — actual vs requested resource usage
→ Network traffic analysis — identify NAT, egress, and cross-region waste
→ Cloud SQL and storage audit — non-production instance sizing
→ Zombie resource cleanup — deprecated projects still incurring cost
→ FinOps ownership structure — budget alerts, labels, cost attribution

Frequently Asked Questions

How do I find out why my GCP bill is increasing?

Enable Cloud Billing export to BigQuery, then query by service and resource over the past 90 days. NAT Gateway, GKE node pools, and Cloud SQL are the most common culprits for unexplained increases.

Buoyant Cloud typically identifies $15,000–$25,000/month in addressable waste on a standard GCP audit for a Series A/B company. The most common sources are over-provisioned GKE, NAT Gateway misrouting, and non-production databases running at production size.

GCP FinOps is the practice of aligning engineering, finance, and operations around cloud cost ownership. On GCP it involves Cloud Billing export, budget alerts per project and service, resource labelling by team, and architecture decisions that make waste structurally impossible rather than just visible.

Buoyant Cloud’s initial GCP cost audit takes 3–5 business days and produces a prioritised list of addressable waste with implementation effort estimates for each item.

Not necessarily. The fastest savings come from architecture fixes — rightsizing GKE, rerouting NAT traffic, and scheduling non-prod databases — not from adding another monitoring tool. Tools help with ongoing visibility; architecture fixes reduce the underlying waste.

Start With a Free GCP Cost Review

If your GCP bill has increased in the last 3 months and your team can't explain the full driver — book a free 30-minute GCP cost review with Amit Malhotra.