Cloud-Native Migration (Stratus)

Multi-year migration from monolithic hub-spoke to stateless Kubernetes platform

Overview

Led the multi-year migration of Appian's cloud infrastructure from a monolithic hub-spoke architecture to a stateless, Kubernetes-based platform (internally codenamed Stratus). This transformation serves 1,200+ enterprise customers across 21 global regions.

Stratus is Appian's next-generation cloud-native platform that enables teams to innovate more quickly, efficiently, and securely than ever before. It provides the foundational infrastructure for all Appian cloud services including AI Skills, Portals, Elastic Process Execution, and Process Mining.

Partnered cross-functionally with engineering, architecture, and InfoSec to meet 25+ compliance frameworks (FedRAMP Moderate, FedRAMP High, SOC 2, HIPAA, PCI-DSS, DoD IL5), while enabling rapid, secure delivery of Appian's AI and automation capabilities.

Results

• Eliminated 7 categories of production outages entirely
• Reduced overall site restart downtime by 23%
• Reduced static content response time by up to 30%
• Cut web server provisioning from 5-30 minutes to seconds
• Enabled new services (Document Extraction improvements) to go from conception to production in under a week
• Met 25+ compliance frameworks including FedRAMP and DoD IL5 preparation
• Expanded to 21 global regions including FedRAMP-certified government cloud
• Also delivered as Appian on Kubernetes for self-managed customers wanting the same cloud-native benefits

My role and the organization behind the migration

I served as Senior Software Engineering Manager and functional group lead from January 2022–March 2023, followed by Director & Group Lead, Cloud Services & Platform Engineering from April–December 2023. I aligned five squads and approximately 50 direct and indirect reports around customer needs, platform priorities, and operational responsibilities.

Working with architecture, engineering, and security, I led the group delivering the Kubernetes web-layer redesign: 23% less restart downtime, up to 30% better static-content response time, and provisioning reduced from 5–30 minutes to seconds. These are outcomes of the cloud migration work, preceding my Agentic AI role.

I established a Cloud Native team in Chennai, hiring six developers and a senior manager and defining its mission and roadmap. Moving incident response to service-team ownership placed subject-matter experts on first response and connected development, maintenance, and operations.

For HITRUST, I led the group whose web-layer controls completed readiness work in one sprint, supporting the broader certification effort. The scope was the web layer, with cross-functional engineering, architecture, and information-security partners.

Key Principles

• Elastic Scale — Containerized architecture enables horizontal pod auto-scaling. Services scale up under load and scale down when idle, ensuring efficient resource utilization. Handles 8.4M+ logins, 2.2B+ workflows, and 5,500+ deployments per quarter.
• Multi-Tenancy with Per-Customer Isolation — Each customer maintains an isolated instance and database. Data is encrypted and belongs only to the customer. Multiple layers of encryption ensure zero knowledge between tenants even on shared infrastructure.
• Zero Trust Security — Hyper least-privilege enforcement via policy and automation. Continuous re-authentication with hardware MFA tokens. Assume-breach network posture with tightly controlled service-to-service communication and minimized blast radius.

Technical Architecture

• Kubernetes Orchestration — Decomposed monolithic services into stateless microservices. Kubernetes reconciliation loop ensures optimal state at all times.
• Containerization — Lightweight packages of application code with dependencies. Dense packing of functionality with less resources, enabling more resiliency than dedicated predecessors.
• Independent Service Scaling — AI Skills, Portals, and Elastic Process Execution built natively in Stratus can scale and deploy independently, enabling fixes and new functionality to ship in under a week without platform version changes or downtime.
• Edge Layer — Request routing and load distribution across application server instances.
• Kafka for High-Throughput Messaging — Scaled independently for Elastic Process Execution workloads to prevent bottlenecks.

Videos

Watch the full Appian World presentation →