Worldly:
Building a Production-Grade DevOps Operating Model on AWS

About the client

Worldly is the world's leading sustainability and supply chain intelligence platform, operating in the ESG technology and sustainability data industry. Worldly provides brands, retailers, manufacturers, and suppliers across the fashion, apparel, footwear, outdoor, home goods, and consumer goods sectors with the tools, data, and insights needed to measure, manage, and improve their environmental and social impact. As the exclusive licensee of the Higg Index, the global standard for sustainability measurement in the apparel and footwear industries, Worldly serves over 40,000 organizations across more than 100 countries.

Worldly's platform is mission-critical for its customers: it underpins ESG reporting workflows, regulatory compliance processes, and supply chain risk assessments that directly influence business decisions and public commitments across the global consumer goods industry.

Teracloud engaged with Worldly's engineering team as a long-term cloud operations partner, supporting the design, implementation, and optimization of their AWS infrastructure and DevOps practices to meet the demands of a rapidly growing, globally distributed platform.

Challenges

Worldly is the world's leading sustainability and supply chain intelligence platform, trusted by 40,000+ brands, retailers, suppliers, and manufacturers across the fashion, outdoor, home goods, and consumer goods industries. As the exclusive licensee of the Higg Index, the most widely adopted sustainability measurement framework in apparel and footwear, Worldly's platform is mission-critical: its customers depend on it to meet ESG reporting obligations, comply with emerging global regulations, and manage environmental and social risks across their supply chains.

In early 2025, Worldly entered a phase of accelerated growth driven by a strategic corporate expansion, including the extension of platform availability to new geographies such as China and Europe. This rapid, multi-dimensional growth created a critical inflection point: the informal DevOps practices that had supported the organization through its earlier stages were no longer sufficient to govern a globally distributed, multi-region, multi-account AWS environment at the required scale.

The core business challenge was the absence of a formalized and scalable DevOps framework governing the full software delivery lifecycle. This manifested in four specific gaps:

Infrastructure changes lacked standardization. While Terraform was in use, there were no enforced governance controls, automated validation gates, or mandatory peer review processes, introducing infrastructure drift risk and increasing the probability of production incidents caused by manual errors.
Observability and incident response were incomplete. Critical infrastructure components lacked comprehensive monitoring. Alert thresholds were uncalibrated and failed to reflect real workload behavior. There was no structured on-call rotation or escalation path. As a result, incidents were frequently detected and reported by end-users before the engineering team was formally alerted.
Disaster Recovery was undocumented and untested. For a platform serving 40,000+ organizations with active ESG reporting obligations, the absence of validated DR processes and documented runbooks represented a direct threat to business continuity.
Access and governance complexity was growing unsustainably. Expansion to new AWS accounts and regions created IAM management challenges that, without formal controls, would lead to over-permissioned roles, auditability gaps, and security exposure.

Prior to the engagement, p95 API latency stood at 2,500ms and Mean Time to Resolution (MTTR) averaged 1 hour 53 minutes, metrics that reflected the operational gaps described above and established the baseline against which improvements were later measured.

Risk of not addressing this challenge

Operational: Platform instability or extended outages affecting 40,000+ global users and their ability to meet ESG regulatory deadlines, directly damaging Worldly's reputation and customer retention in a trust-dependent market.

Expansion: Without repeatable and auditable infrastructure practices, safe deployment to new regions (China, Europe) would remain high-risk and block Worldly's growth roadmap.

Security and compliance: An ungoverned multi-account AWS environment at scale exposes the company to unauthorized access, data breaches, and audit failures, incompatible with the trust enterprise customers place in a platform handling sensitive supply chain and ESG data.

Competitive: In a rapidly growing, regulation-driven ESG intelligence market, platform reliability and scalability are key differentiators. Failing to address these gaps would undermine Worldly's position as the market's most trusted platform.

Goals and Objectives

Working backwards from the operational and business outcomes required to support Worldly's growth phase, Teracloud and Worldly jointly defined the following goals and objectives as part of the engagement.

Business Objectives

Ensure platform reliability at scale. Maintain service levels aligned to enterprise ESG reporting requirements for 40,000+ global users, minimizing any disruption that could impact customers' regulatory compliance workflows.

Enable safe multi-region expansion. Support Worldly's strategic expansion to China and Europe with the confidence that new workload deployments would be governed, repeatable, and carry minimal business continuity risk.

Reduce incident response time. Formalize detection and escalation mechanisms to significantly reduce Mean Time to Acknowledge (MTTA) and Mean Time to Resolution (MTTR), shifting from reactive to proactive operations.

Establish enterprise-grade cloud governance. Build an auditable, secure, and compliant AWS operating model capable of supporting Worldly's enterprise customer expectations and growing regulatory scrutiny in the ESG space.

Technical Objectives

Formalize Infrastructure as Code governance. Enforce Terraform-based IaC with mandatory peer review, automated validation (plan and apply), and full version control via GitHub, ensuring all infrastructure changes are traceable, repeatable, and production-safe.

Implement a structured CI/CD pipeline. Build and standardize the software delivery pipeline covering code review, automated unit and integration testing (Jest + AWS CodeBuild), artifact management (Amazon ECR), and controlled promotion across Development, Staging, and Production environments via Jenkins.

Deploy end-to-end observability. Implement Datadog APM for distributed tracing and application performance monitoring, combined with infrastructure-level monitoring, calibrated alerting, and PagerDuty-based on-call management and escalation policies.

Define and validate a Disaster Recovery strategy. Design, document, and test a Pilot Light DR architecture in a secondary AWS region (us-east-1), with clearly defined RTO/RPO targets and runbooks covering all core recovery scenarios.

Establish multi-account governance. Implement AWS Organizations with dedicated accounts per environment and function, centralized identity management via IAM Identity Center, and Service Control Policies (SCPs) to enforce security baselines across all accounts.

Align with the AWS Well-Architected Framework. Adopt AWS best practices across the Security, Operational Excellence, and Reliability pillars, systematically identifying and remediating risks across the platform architecture.

Solution and Execution

Teracloud partnered with Worldly to design and implement a comprehensive DevOps operating model across four interconnected workstreams:

Infrastructure as Code Governance: Teracloud standardized all AWS infrastructure provisioning using Terraform, with the full codebase stored in GitHub and subject to mandatory peer review via pull requests. Every infrastructure change, from VPC modifications to ECS task definitions, requires a reviewed and approved terraform plan before execution. This process eliminated manual configuration and infrastructure drift, and established a fully auditable trail of every change made to Worldly's production environment. The same IaC approach governs all AWS accounts (Production, Data, Security, DR) and all regions (us-west-2 primary, us-east-1 DR, China), ensuring consistency and repeatability across the entire multi-account architecture.

CI/CD Pipeline: Teracloud designed and implemented Worldly's end-to-end software delivery pipeline, spanning from code commit to production deployment. Pull requests trigger automated test execution in AWS CodeBuild, covering unit tests (Jest, sharded 3-way), integration tests against CouchDB and Elasticsearch, and infrastructure validation (terraform validate + terraform plan). Security vulnerability scanning runs daily via Dependabot, with results surfaced to the engineering team via Slack. Validated Docker images are built and published to Amazon ECR using immutable, commit-based tags. Controlled promotion across Development, Staging, and Production is managed via Jenkins, with production releases gated on successful automated validation in Staging and a formal Deployment Readiness Checklist review. Post-deployment, automated ECS stability checks and build version assertions confirm successful delivery before the pipeline completes.

Observability and Incident Management: To address Worldly's incomplete monitoring and lack of structured incident response, Teracloud deployed a full observability stack. Datadog APM was instrumented across Worldly's application services, providing distributed tracing, application performance monitoring, and service-level dashboards. Infrastructure metrics and logs were consolidated, and alerting thresholds were calibrated to real workload behavior, replacing generic, noisy alerts with actionable signals correlated to business impact. PagerDuty was configured with structured on-call rotations, escalation policies, and incident lifecycle management, formalizing a process that had previously been ad hoc. The combination reduced MTTR by 35% and shifted the team from reactive to proactive operations.

Disaster Recovery: Teracloud designed and implemented a Pilot Light DR architecture in a secondary AWS region (us-east-1), with the DR infrastructure maintained in Terraform to ensure full parity with production. Recovery procedures were documented in a formal DR runbook and validated through a tabletop DR exercise, with the DR environment formally certified against Worldly's Business Continuity Plan. RTO (24 hours) and RPO (36 hours) targets are contractually committed in Worldly's SLA and were discussed and agreed upon during solution architecture workshops conducted by Teracloud.

Cloud Security and Account Governance: Teracloud established a multi-account architecture under AWS Organizations, with dedicated accounts for Production, Data, Security, and Disaster Recovery. Service Control Policies (SCPs) enforce security baselines across all accounts, preventing privilege escalation and ensuring consistent policy application. Centralized identity management was implemented via AWS IAM Identity Center, with least-privilege role assignments for all human and service identities. Continuous security posture monitoring is provided by Amazon GuardDuty, AWS Security Hub, and Amazon Inspector, with findings centralized in the Security account. AWS CloudTrail provides immutable API audit logs across all accounts and regions.

Architecture Overview

Worldly's platform runs on AWS across a multi-account, multi-region architecture governed under AWS Organizations. The account structure separates workloads by function and risk boundary: dedicated accounts for Production, Data, Security, and Disaster Recovery, each governed by Service Control Policies (SCPs) enforced at the organizational level.

Worldly AWS Architecture, Production (us-west-2), Data Account, and DR (us-east-1)

Production Account: us-west-2: The production environment is built on a three-tier VPC architecture spanning two Availability Zones (us-west-2a, us-west-2b), with network segmentation enforced via Security Groups and Network ACLs. Inbound traffic is handled via an Application Load Balancer, fronted by Amazon CloudFront, AWS WAF, and AWS Certificate Manager (ACM). DNS routing is managed via Amazon Route 53. Static assets are served via Amazon S3. Application workloads run as containerized services on Amazon ECS (EC2 launch type) across private subnets in both AZs.
The application database layer consists of CouchDB on EC2, with a primary instance in us-west-2a and a standby replica in us-west-2b, and Amazon Elasticsearch deployed across both AZs for the advanced query layer. Automated backups are managed via AWS Backup. Infrastructure and instance management is supported by AWS Systems Manager (SSM).
Security tooling runs continuously across the account: Amazon GuardDuty (threat detection), AWS Security Hub (centralized findings), Amazon Inspector (vulnerability scanning), AWS CloudTrail (API audit logging), AWS Config (configuration compliance), and Amazon CloudWatch (metrics and alarms).

Data Account: us-west-2: The data warehouse layer runs in a dedicated Data Account, isolated from the production workload via a separate VPC connected through VPC Peering. Amazon RDS is deployed in a multi-AZ configuration (primary in us-west-2a, replica in us-west-2b) for analytical workloads. The account includes its own security and monitoring stack (GuardDuty, CloudTrail, Security Hub, Inspector, Config, CloudWatch) and AWS Backup for automated data protection.

Software Delivery Pipeline: Infrastructure changes flow through Terraform with mandatory GitHub pull request review and automated plan validation before apply. Application changes flow through: GitHub → AWS CodeBuild (automated testing and Docker image build) → Amazon ECR (immutable image storage) → Jenkins (controlled promotion across Development, Staging, and Production). All container images are tagged with commit hashes, guaranteeing artifact immutability across environments.

DR Account: us-east-1 (Pilot Light): A Pilot Light DR environment is maintained in us-east-1, built and managed using the same Terraform codebase as production to prevent infrastructure drift. The DR account mirrors the production architecture: CloudFront, WAF, ACM, Route 53, ALB, ECS cluster (scaled to zero under normal conditions), CouchDB (main + replica), Elasticsearch, ECR, CodeBuild, S3, and the full security tooling stack. CouchDB data is continuously replicated from the primary region via VPC Peering. In the event of a regional failure, Route 53 routing is updated to direct traffic to the DR environment, and ECS workloads are scaled up from standby. The DR environment has been formally certified and validated through a tabletop exercise against Worldly's Business Continuity Plan.

KPIs And Outcomes

The following key performance indicators were measured against baseline values captured prior to the engagement, reflecting the operational improvements achieved through Teracloud's DevOps implementation.

p95 API Latency (76.5% Reduction): End-to-end application performance improvements, including containerization, multi-AZ load balancing, and Datadog APM-driven identification of latency bottlenecks, reduced the p95 API response time from 2,500ms to 587ms. This directly impacts the experience of 40,000+ users relying on the platform for time-sensitive ESG reporting workflows.

Platform SLO (99.809% Availability): The formalization of the CI/CD pipeline, multi-AZ deployment, automated deployment readiness checks, and structured incident management via PagerDuty enabled Worldly to achieve and sustain a measured platform availability of 99.809% in production, meeting enterprise SLA expectations for a mission-critical ESG compliance platform.

MTTR (35% Reduction): The deployment of Datadog APM with calibrated alerting and PagerDuty with structured on-call rotations transformed incident response from a reactive, ad hoc process to a managed, escalation-driven workflow. MTTR dropped from 1 hour 53 minutes to 1 hour 14 minutes, reducing the business impact of production incidents on Worldly's customers.

Business Impact

The operational improvements delivered through this engagement translate directly into Worldly's capacity to grow, compete, and fulfill its market mission.

A platform that now operates at 99.809% availability with a p95 latency under 600ms is a fundamentally different commercial proposition than one where performance was unmeasured and incident response was ad hoc. For Worldly's enterprise customers, global brands and retailers with active ESG reporting obligations and regulatory deadlines, platform reliability is not a technical attribute; it is a contractual expectation and a condition of trust. The ability to commit to, measure, and evidence SLOs strengthens Worldly's positioning in enterprise sales cycles and renewal conversations.

The 76.5% reduction in API latency and the 35% reduction in MTTR also directly reduce the cost of operations: fewer escalations, faster recovery, and less engineering time spent on reactive firefighting, freeing capacity for product development and platform expansion. Combined with a governed, repeatable deployment process, Worldly's engineering team can now ship changes to production with confidence and at the cadence that a fast-growing platform demands.

Perhaps most significantly, the DevOps foundation established through this engagement is what made Worldly's geographic expansion viable. Safe, auditable, and consistent infrastructure deployment across new regions, China and Europe, requires exactly the governance model that Teracloud designed and implemented. Without it, expansion would carry unacceptable operational and security risk for a platform that serves as the global ESG compliance backbone for tens of thousands of organizations.

Get in touch

Take your business to the next level.

Worldly: Building a Production-Grade DevOps Operating Model on AWS