How Ideas2IT Cut Cognetivity's AWS Infrastructure Costs by 60% Through a Full-Stack Cloud Audit and Compliance-Driven Remediation

Cognetivity's AWS environment had grown faster than it was governed, with idle resources, unreviewed spend, and security exposure accumulating across UK and UAE clusters. Ideas2IT audited the full infrastructure stack, removed dormant compute, remediated the security findings, and restructured the environment to scale without the overhead.

Client

Cognetivity Neurosciences

Industry

All Industries

Service

Cloud

Compliance

HIPAA

Stack

AWS · EKS · Terraform · Argo CD

01 Challenge

Cognetivity's CognICA platform ran across UK and UAE AWS clusters with no unified governance layer. Dormant compute sat unreviewed, infrastructure spend had no right-sizing discipline, and security posture had not been validated against the compliance requirements of regulated healthcare markets.

02 Solution

Ideas2IT ran a full AWS infrastructure audit using Cost Explorer, Trusted Advisor, and the Well-Architected Framework. Dormant EKS clusters and over-provisioned RDS instances were identified and removed. VPC peering was corrected across dev, QA, and production environments. Security controls, IAM policies, and CI/CD pipelines were remediated and Argo CD was layered into GitLab CI/CD for deployment monitoring.

03 Outcome

Infrastructure costs dropped 60% [NEEDS CLIENT INPUT]. Dormant cluster overhead was eliminated across UK and UAE environments. Security posture was brought into compliance. A Cucumber-based automation framework cut regression testing time and integrated into the CI/CD pipeline with nightly execution.

Phase 01

From unreviewed multi-region sprawl to a right-sized AWS environment

Infrastructure Audit and Dormant Resource Elimination: reducing cloud spend by removing compute that was never in use

The first problem was visibility. Cognetivity's AWS estate had grown across UK and UAE regions without a systematic review of what was actively serving traffic and what had been left running. Ideas2IT ran a full audit using AWS Cost Explorer, Trusted Advisor, and the Well-Architected Framework, mapping every active and dormant resource across EKS clusters, RDS instances, and VPC configurations. Dormant clusters were decommissioned. RDS instances were right-sized. VPC peering connections between EKS and RDS were corrected in Dev and QA UAE environments after a database connectivity fault was identified. The RDS migration completed in both QA UAE and QA UK environments without data loss, and the full production UAE environment came online with EKS, RDS, and data migration completed cleanly.

This Phase Produced

  • AWS infrastructure audit report
  • Cost Explorer, Trusted Advisor, and Well-Architected analysis across UK and UAE environments
  • Dormant EKS cluster decommissioning
  • Removal of idle compute across UAE and UK regions
  • VPC peering remediation
  • Corrected EKS-to-RDS connectivity in Dev and QA UAE environments
  • RDS right-sizing and migration
  • Completed in QA UAE, QA UK, and Production UAE environments
  • Terraform infrastructure-as-code
  • RDS provisioning in London CFC dev cluster via Terraform
  • WIKI documentation
  • Step-by-step VPC peering reference for future cluster creation

Phase 02

Closing the security gaps that a growing AWS estate had left open

Security Posture Remediation and CI/CD Hardening: bringing a regulated healthcare platform into compliance across all environments

With the cost baseline established, the audit surfaced the security findings. IAM access controls were reviewed and tightened. Argo CD was layered into the GitLab CI/CD pipeline as the deployment and health monitoring layer, giving the team continuous visibility into environment state across UK and UAE clusters. The CognICA portal and client application were put through structured test certification: 144 test cases for the portal and 112 for the client application, covering UK and UAE configurations, with discrepancies between requirements and test cases resolved before sign-off. A user action audit API validation completed in UAE LDS. SageMaker replaced Amazon ML across the platform, providing a supported, API-consistent ML infrastructure for multi-region operations and reducing dependency on a deprecated service.

This Phase Produced

  • Argo CD deployment layer
  • Integrated into GitLab CI/CD for deployment and environment health monitoring
  • Portal test certification
  • 144 test cases executed across UK and UAE environments
  • Client app test certification
  • 112 test cases covering Arabic bug fixes across UK and UAE
  • User action audit API validation
  • Completed in UAE LDS environment
  • SageMaker migration
  • Amazon ML replaced with SageMaker across multi-region ML workflows
  • IAM and access control review
  • Security posture remediated across all AWS environments

Phase 03

From manual testing to a scheduled, CI/CD-integrated automation layer

Test Automation Framework and Mobile Platform Stability: eliminating manual regression cycles across a multi-region clinical application

Testing had been entirely manual across Cognetivity's UK and UAE environments. Ideas2IT built an automation framework from scratch on Cucumber, designed it to cover all existing test scenarios, and got client sign-off before integration. The framework was connected to the CI/CD pipeline, nightly execution was scheduled, and Slack notifications were configured so test results were available to the team without manual reporting. On the mobile side, more than 10 bugs were resolved across Android and iOS in the KMM codebase, covering broken navigation flows and missing action events. Apple Silicon compatibility was added to a codebase that had only supported Mac Intel. A region selector feature was built under the login screen for the QA environment. The cancel feature was implemented in Test History on iOS CognICA Live.

This Phase Produced

  • Cucumber automation framework
  • Built from scratch, accepted by client, covering all existing test scenarios
  • CI/CD pipeline integration
  • Nightly test execution scheduled and connected to pipeline
  • Slack test result notifications
  • Automated communication of test execution results to the team
  • KMM bug resolution
  • 10+ bugs fixed across Android and iOS covering navigation and action events
  • Apple Silicon compatibility
  • Updated KMM iOS codebase from Mac Intel-only to Apple Silicon support
  • Region selector feature
  • Built for QA login screen with UI and API integration
  • iOS cancel feature
  • Cancel action implemented in Test History screen, CognICA Live

The Outcome

A 60% reduction in AWS spend and a regulated platform brought into compliance without a production disruption.

Category Metric Description
Infrastructure costs 60% reduction Dormant cluster elimination and right-sizing across UK and UAE environments
Test coverage 256
Test cases certified		 144 portal and 112 client app test cases executed across UK and UAE
Security Remediated IAM, VPC, and CI/CD controls brought into compliance across all environments
ML infrastructure Complete Amazon ML replaced with SageMaker for multi-region, community-supported ML operations
Mobile quality 10+
Bugs resolved		 Navigation, action event, and platform compatibility issues in KMM iOS and Android
Automation Eliminated
Regression cycle		 Cucumber framework replaced manual testing with CI/CD-scheduled nightly execution
The cost reduction followed directly from the audit methodology, not from infrastructure optimisation applied on top of an unreviewed environment. Dormant clusters cannot be right-sized; they have to be found and removed first. The same discipline that produced the cost outcomes drove the security remediation: the Well-Architected review surfaces cost problems and security exposure in the same pass. A regulated healthcare platform operating in UK and UAE markets had both addressed in a single engagement.