PACE: Aviation Carbon Intelligence Platform Delivers Real-Time Emissions Tracking Across 99% of Global Flights
Cloud-native Kubernetes architecture enables multi-tenant SaaS platform serving airports, airlines, and financial institutions with zero-downtime cloud migration.
Executive Summary
Fexco is a global financial services and business solutions provider operating across 29 countries, delivering foreign exchange, corporate payments, tax-free retail services, and innovative technology platforms for regulated industries.
The Problem
Fexco needed to launch PACE; a cloud-native aviation carbon intelligence platform capable of processing emissions data from 99% of worldwide flights in real-time, serving multiple client types (airports, aircraft portfolio owners, financial institutions) with distinct feature requirements – all from a single, scalable SaaS architecture
The Solution
Zartis designed and built the entire PACE platform from the ground up. We delivered a cloud-native, Kubernetes-based multi-tenant system with full-stack application development, infrastructure provisioning, monitoring, ETL pipelines, and seamless cloud vendor migration capabilities.
The Outcomes
99%+ Global Flight Coverage: Real-time emissions data processing across 370,000+ airline routes with 24-hour data refresh cycles
Zero-Downtime Cloud Migration: Successfully transitioned between cloud providers (Azure to GCP) in 4-6 months without service disruption
Multi-Tenant Architecture at Scale: Single platform serving airports, portfolio owners, and investors with role-based feature access and unified data governance
Significant BI Cost Reduction: Migration from Azure Synapse to GCP BigQuery delivered same functionality at a fraction of operational costs
Production-Grade Observability: Distributed tracing with Grafana stack (Prometheus, Loki, Tempo) plus automated synthetic monitoring across critical user journeys
Fexco provides financial services ranging from dynamic currency conversion and corporate payments to innovative technology platforms. Beyond traditional financial services, Fexco develops domain-specific software solutions for regulated industries including aviation, treasury management, and cross-border payments.
About PACE:
PACE is Fexco’s aviation carbon intelligence platform,; a software suite designed for the aviation and finance industries focused on calculating emissions and developing data-driven strategies for carbon reduction. The platform serves airports optimising aircraft operations, fleet owners managing portfolios, and financial institutions tracking investment sustainability.
The Business Opportunity
The aviation industry faces mounting regulatory pressure to reduce carbon emissions whilst investors and financial institutions need transparent ESG data for portfolio risk assessment. Airports, airlines, and aircraft portfolio owners require granular, real-time emissions intelligence to meet compliance requirements, optimise operations, and reduce costs.
Fexco identified a critical market opportunity:
Build a platform that serves multiple stakeholders – airports tracking operational emissions, portfolio owners optimising fleets, and financial institutions assessing investment risk – from a single source of truth for aviation emissions data.
The Technical Challenge
Fexco needed help to cope with these demainding technical requirements:
Real-time Data Processing
Capture and process emissions data from 99% of worldwide flights with 24-hour refresh cycles
Multi-Tenancy at Scale
Serve airports, airlines, portfolio owners, and investors with distinct feature sets and data access patterns from a single platform
Granular Data Fidelity
Track emissions per aircraft, per flight, with market information linked to portfolios, fleets, airports, and client accounts
Role-Based Access
Airport operators need airport-specific data; portfolio owners need fleet performance; investors need aggregated portfolio emissions
Scalability and Performance
Handle massive data volumes whilst maintaining sub-second query response times for interactive dashboards
Why Traditional Approaches Weren't Working
Cloud-native from day one was non-negotiable
The platform’s multi-tenant architecture, where multiple clients access different features and data views from shared infrastructure, made traditional monolithic architecture infeasible. Building separate deployments per client would create maintenance nightmares and prevent the economies of scale required for SaaS economics.
Product-focused vs. bespoke development
Fexco needed a unified software platform serving multiple use cases, not custom one-off implementations per client. Airport operators, aircraft portfolio owners, and institutional investors (banks, corporations) have distinct needs, but all require access to the same real-time flight and emissions data. A shared architecture with centralised data governance and flexible feature access was essential.
Vendor flexibility required
The platform needed to remain cloud-agnostic to avoid vendor lock-in and capitalise on best-in-class BI tooling across providers. This meant Kubernetes-based architecture from the start, enabling seamless migration between cloud vendors without application rewrites.
Internal capacity constraints
Building a cloud-native, multi-tenant platform with real-time data pipelines, modern frontend, and production-grade observability required expertise Fexco didn’t have in-house. They needed a partner who could both design the architecture and execute the full stack, not just provide staff augmentation.
Why zartis
The Zartis Approach
Zartis wasn't brought in to augment an existing team; we built the PACE platform from scratch, owning the entire technical stack end-to-end.
Full Ownership from Day One:
Zartis took responsibility for the complete platform; application business logic, infrastructure provisioning, external service integrations, monitoring and alerting, ETL processes, and operational management. This wasn't staff augmentation; it was full-stack product development with Zartis engineers acting as the core technical team.
Embedded Product Team:
Zartis engineers worked directly with Fexco's product and business stakeholders, translating aviation domain requirements and financial use cases into technical architecture decisions. Our team participated in product planning, sprint ceremonies, and technical roadmap prioritisation.
MVP-Driven Feature Delivery:
The platform launched with core functionality and incrementally added major features (Airports module, Portfolio Infographics, Refleet) using feature flags. This allowed controlled rollouts and early customer feedback without destabilising the platform.
Proactive Technical Debt Management:
Despite aggressive feature delivery, code quality remained top priority. Automated testing, QA processes, and a dedicated tech debt backlog ensured the platform remained maintainable as complexity grew.
Our approach
What We’ve Built & Delivered
Cloud-Native Kubernetes Platform Architecture
Zartis designed and implemented the complete infrastructure from scratch:
Key Deliverables
Kubernetes-Based Microservices
Application services containerised and orchestrated for scalability, resilience, and cloud portability
Multi-Cloud Deployment
Initially on Azure, later migrated to GCP - architecture design enabled seamless transition
GitOps Deployment Model
ArgoCD for declarative infrastructure and application management, ensuring consistent deployments across environments
CI/CD Automation
Jenkins pipelines for automated builds, testing, and deployments
Why This Mattered
The Kubernetes foundation enabled the cloud vendor migration without application rewrites. When Fexco negotiated better commercial terms with GCP, the platform transitioned in months, not years - without user impact.
Full-Stack Application Development
Zartis built both frontend and backend from the ground up.
Frontend
Angular-based web application with TypeScript
Implementation of web widgets with StencilJS, as well as a WordPress plugin for it
Interactive data visualisation dashboards tailored to different user roles (airports, portfolio owners, investors)
Backend
Golang microservices handling business logic, API orchestration, and real-time data processing
BigQuery as cloud data warehouse and Looker as Business Intelligence / data processor (migrated from Elasticsearch
PostgreSQL for transactional data and MongoDB for document storage
Pub/Sub messaging for event-driven architecture
Temporal Workflows for complex, long-running business processes (ETL jobs, data synchronisation)
Why This Mattered
Building with modern, cloud-native technologies from day one meant the platform could scale horizontally as data volumes grew. The event-driven architecture enabled real-time data processing - critical for 24-hour refresh cycles across 370,000+ airline routes.
Business Intelligence & Data Pipeline
PACE’s value proposition depends on transforming raw flight data into actionable emissions intelligence. Initially, the platform used Elasticsearch for data aggregation and search. Zartis migrated it to Looker for all data transformations, achieving a unified data source and cost reduction.
Key Deliverables
BigQuery Data Warehouse
Centralised analytics layer for complex queries across massive datasets
Looker Integration
BI platform providing interactive dashboards and self-service analytics for end users
ETL Pipelines
Automated data ingestion, transformation, and loading from multiple aviation data sources
Data Consistency Architecture
Looker API as single source of truth, eliminating data inconsistencies across platform modules
Why This Mattered
Portfolio owners and investors need confidence in emissions calculations, as inconsistent data across dashboards would destroy trust. Centralising on Looker created a single source of truth whilst reducing infrastructure costs.
Production-Grade Observability & Monitoring
We implemented comprehensive monitoring from day one, not as an afterthought.
Grafana Stack Deployment:
- Prometheus: Metrics collection for application performance, resource utilisation, and business KPIs
- Loki: Log aggregation for debugging and incident analysis
- Tempo: Distributed tracing across microservices, enabling rapid bottleneck identification
Automated Synthetic Monitoring:
Rather than waiting for users to report issues, Zartis built automated testing simulating real user journeys:
- Synthetic browser sessions login, execute critical features (data queries, portfolio calculations, emissions reports), verify results, and logout
- Immediate alerting on failures or performance degradation
- Continuous validation of core user workflows across deployments
Integrated Alerting:
Monitoring alerts elevate issues to DevOps, development, and product teams via Microsoft Teams integration, ensuring rapid response to production incidents.
Why This Mattered
With clients relying on PACE for regulatory compliance and operational decisions, downtime or data inaccuracies create significant business risk. Proactive monitoring catches issues before they impact users.
Security & Vulnerability Management
Zartis implemented automated security practices integrated into the development lifecycle.
Key Deliverables
CI/CD Security Scanning
Automated tools analyse code and dependencies for known vulnerabilities during build pipelines
Developer Alerts
Engineers notified immediately when vulnerabilities detected in libraries or dependencies
Proactive Patching
Automated update process ensures timely fixes applied to both application code and third-party libraries
Why This Mattered
Aviation and financial services clients require robust security. Embedding vulnerability detection into CI/CD prevents security debt accumulation and ensures compliance with industry standards.
Our approach
Key Technical Decisions & Trade-Offs
Kubernetes-Native Architecture for Cloud Portability
We built on Kubernetes from day one rather than using cloud-specific services (AWS ECS, Azure Container Instances).
Why This Mattered:
When Fexco negotiated better commercial terms with GCP, including superior BI tooling (BigQuery + Looker vs. Azure Synapse), the migration happened in 4-6 months with zero user impact. The application synchronised smoothly using ArgoCD’s GitOps approach. Database migration occurred during a short maintenance window outside business hours.
Technical Tade-off & Business Impact:
Kubernetes adds operational complexity compared to serverless or managed container services. However, this investment paid off immediately when the business needed cloud flexibility. We avoided vendor lock-in whilst capitalising on best-in-class BI capabilities and cost savings from competitive cloud pricing.
Event-Driven Architecture with Pub/Sub Messaging
We used Google Cloud Pub/Sub for asynchronous communication between microservices rather than synchronous API calls.
Why This Mattered:
PACE processes massive data volumes; flight telemetry, emissions calculations, portfolio updates in real-time. Synchronous architectures would create bottlenecks and cascade failures. Pub/Sub enables independent scaling of data ingestion, processing, and presentation layers.
Business Impact:
When new flight data arrives, it triggers events consumed by multiple services (emissions calculation, portfolio updates, alerting) without blocking the ingestion pipeline. The platform maintains 24-hour data refresh cycles across 370,000+ routes without performance degradation, even as client volumes grow.
Temporal Workflows for Complex Business Processes
We used Temporal for orchestrating long-running, stateful workflows (ETL jobs, multi-step data transformations, external API integrations).
Why This Mattered:
Aviation data flows through multiple transformation stages: raw telemetry → emissions calculations → portfolio aggregations → financial reporting. Traditional cron jobs or state machines fail under complexity. Temporal provides durable execution with automatic retries, failure recovery, and visibility into workflow state.
Business Impact:
Data pipelines run reliably without manual intervention. When failures occur, workflows automatically resume from the last successful step—preventing data loss and reducing operational burden.
Eliminating Elasticsearch for Cost Optimisation
Our team replaced Elasticsearch with Looker API for all data aggregations and transformations.
Why This Mattered:
- Consistency: Looker became the single source of truth, eliminating data discrepancies across platform modules
- Cost reduction: Elasticsearch demanded significant compute resources; Looker delivered same analytical capabilities at lower operational cost
Business Impact:
This required refactoring data access patterns across the frontend, but delivered immediate ROI through reduced infrastructure spend whilst improving data reliability. We lowered operational costs whilst increasing user trust through consistent data across all dashboards and reports.
Feature Flags for Risk Mitigation
We deployed major features (Airports module, Portfolio Infographics, Refleet) behind feature flags rather than “big bang” releases.
Why This Mattered:
Different client types need different feature sets. Feature flags enabled:
- Gradual rollouts: Test features with subset of users before full release
- Client-specific configurations: Enable advanced features for enterprise clients whilst maintaining simpler experience for others
- Rollback capability: Disable features instantly if issues arise without redeploying code
Business Impact:
Faster feature delivery with lower risk. The product team could iterate based on real user feedback before committing features to all clients.
GitOps with ArgoCD for Deployment Consistency
Our team decided to use ArgoCD to manage application deployments via Git as the source of truth.
Why This Mattered:
With multiple environments (dev, staging, production) and frequent deployments, manual infrastructure changes create configuration drift. GitOps ensures:
- Declarative infrastructure: All configuration defined in Git repositories
- Automated synchronisation: ArgoCD detects drift and automatically reconciles actual state with desired state
- Audit trail: Every infrastructure change tracked in Git history
Business Impact:
Cloud migration synchronised smoothly because ArgoCD handled the declarative deployment process. No manual intervention required, eliminating human error.
the results
Outcomes & Measurable Impact
Platform Scale & Performance
Three Strategic Platforms Delivered
Successfully developed and deployed three mission-critical platforms supporting UK energy transition—grid flexibility management, market settlement, and enterprise integration
Platform processes emissions data from virtually all commercial flights worldwide
Comprehensive coverage enabling portfolio-level and airport-specific emissions analysis
Real-time intelligence for operational and compliance use cases
Thousands of Aircraft Portfolios
Multiple clients managing extensive fleet data through the platform
Sub-Second Dashboard Queries
Interactive data visualisation maintained despite massive underlying datasets
Cloud Migration & Cost Efficiency
Zero-Downtime Cloud Migration
Azure to GCP transition completed in 4-6 months without service interruption
Significant BI Cost Reduction
GCP BigQuery delivered same analytical capabilities as Azure Synapse at fraction of operational cost
Infrastructure Optimisation
Elasticsearch removal reduced compute costs whilst improving data consistency
Database migration in maintenance window
Short outage window outside business hours ensured no user impact
Platform Reliability & Quality
Distributed Tracing with Grafana Stack
Prometheus, Loki, and Tempo enable rapid performance bottleneck identification
Automated Synthetic Monitoring
Continuous validation of critical user journeys (login → feature execution → logout) catches issues before users report them
Integrated Alerting System
DevOps, development, and product teams notified immediately via Microsoft Teams when issues detected
Proactive Security Posture
Automated vulnerability scanning in CI/CD pipelines prevents security debt accumulation
Development Velocity & Maintainability
MVP-Driven Feature Delivery
Platform launched with core functionality, then iteratively added major features (Airports, Portfolio Infographics, Refleet) using feature flags
Tech Debt Backlog Maintained
Despite aggressive feature delivery, code quality remained top priority with automated testing and QA processes
Feature Flag Strategy
Major features deployed behind flags, enabling gradual rollouts and instant rollback capability without code redeployments
Automated Testing Coverage
Testing strategy minimised bugs and production incidents whilst maintaining rapid iteration velocity
Multi-Tenant Architecture Success
Role-Based Feature Access
Airport operators, portfolio owners, and investors access distinct feature sets from single platform
Unified Data Governance:
Single source of truth (Looker) ensures consistent data across all client views
Scalable Multi-Tenancy
Architecture supports growing client base without performance degradation
What this enabled next
Strategic Value Beyond Initial Build
PACE’s granular emissions tracking positions Fexco’s clients to meet evolving aviation carbon regulations, including EU Emissions Trading System (EU ETS) and Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA). The platform’s audit trail and data fidelity meet regulatory reporting requirements.
Financial institutions (banks, asset managers, corporations) can now assess portfolio-level emissions risk and track sustainability commitments. This addresses growing demand for transparent ESG data in investment decision-making.
Airlines and airports use PACE to identify fuel inefficiencies and operational patterns driving unnecessary emissions, enabling simultaneous cost reduction and sustainability improvements. Data visualisation highlights opportunities for fleet optimisation and route efficiency gains.
Client Relationship Evolution
What began as building a single aviation carbon platform has evolved into a strategic technology partnership. Zartis now maintains ongoing engagement with Fexco across multiple product lines, demonstrating sustained value delivery:
- PACE platform: Continued feature development, infrastructure management, security remediation, and operational monitoring
- Fusion (Foreign Currency Exchange): Cross-border FX application for retail currency sales
- Delivery Platform: DevOps infrastructure providing CI/CD across all Fexco IT projects
- Fexco Smart Assist: Azure OpenAI-based internal documentation assistant
Building Cloud-Native Platforms for Regulated Industries?
If your organisation needs to launch a multi-tenant SaaS platform in aviation, finance, or other regulated industries with demanding requirements for data fidelity, compliance, and real-time intelligence, we combine strategic architecture expertise with full-stack execution capabilities.