40% lower AWS spend in a quarter — while shipping twice as often.

Meridian is a payments platform processing millions of transactions a day. Growth had been kind to the product and brutal to the AWS bill — spend was climbing 12% month over month, deploys were slow and nervous, and a self-managed Kubernetes estate had become a second full-time job for the platform team.

The problem

By the time Meridian called us, three things were true at once: nobody could forecast next month’s bill, the platform team was spending more time babysitting clusters than shipping, and an upcoming SOC 2 audit was looming with no clear owner.

• AWS spend growing 12% MoM with no clear driver.
• Self-managed Kubernetes consuming ~60% of platform-team time.
• Deploys gated on manual steps; releases happened weekly at best.
• SOC 2 controls partially in place, undocumented.

What we did

We opened with our standard two-week assessment, then sequenced the work into reversible waves so the product team never had to stop.

1 · Re-platform off self-managed Kubernetes

We moved stateless services to Amazon ECS on Fargate and migrated the primary datastore from a self-run Postgres to Aurora PostgreSQL. That alone removed the cluster-babysitting tax and unlocked autoscaling that actually tracked traffic.

2 · A FinOps program, not a one-time cleanup

We instrumented every account, killed idle and over-provisioned resources, and built a Savings Plans strategy covering 82% of steady-state compute. Crucially, we stood up showback so each team could see — and own — its own spend.

“For the first time, finance and engineering were looking at the same number and agreeing on what to do about it.”

3 · Golden-path delivery + audit readiness

We codified the landing zone in Terraform, added a CI/CD pipeline with automated checks, and mapped existing controls to SOC 2 — closing the gaps as preventive guardrails rather than documentation debt.

The results

Within a single quarter, the bill was down 40% and predictable, deploy frequency had doubled, and the platform team was back to building product. The SOC 2 audit passed with zero high-risk findings outstanding.

• 40% reduction in monthly AWS spend, now forecastable within 3%.
• 2× deploy frequency with a fraction of the manual steps.
• 99.99% uptime sustained through and after the migration.
• SOC 2 passed with no high-risk findings left open.

Meridian now runs day-to-day on its own, with Enkompass on a light managed retainer for monthly cost and architecture reviews.