How SquareOps Built and Runs QWEEN Jewels' D2C Platform on AWS ECS With 24×7 Managed Operations

Client Overview

QWEEN Jewels is an experiential fine jewellery brand headquartered in Bengaluru, selling certified natural diamond and gemstone jewellery through its own storefront, mobile apps and self-browsing flagship stores. SquareOps was engaged in September 2025, while the company was still operating under its pre-launch name Sucimani, to build the production platform for launch and then run it around the clock under managed AWS operations.

The Challenge

QWEEN was launching a consumer brand on a platform of seventeen microservices and needed it production-grade on day one, without hiring a cloud operations team.

• No operating model. Web frontends, an API gateway, authentication and a dozen backend services for commerce, fulfilment and search needed orchestration, per-service pipelines and health checks. None of it existed yet.
• Downtime costs sales. A failed checkout on a fine jewellery storefront is an abandoned high-value order, so the business wanted a 99.9% uptime commitment backed by an SLA and issues triaged by priority.
• Security and cost from launch. Payments and customer data meant aligning with the AWS Startup Security Baseline before go-live, and a pre-revenue brand cannot carry an oversized AWS bill.

What SquareOps Built

A fixed-scope build between October and November 2025: staging first for the client's QA team, then production, then the storefront went live.

• Container platform. All seventeen services on Amazon ECS with AWS Fargate across multiple availability zones behind an Application Load Balancer, with separate accounts and VPCs for staging and production.
• Managed data services. Fully managed AWS database, caching, streaming, storage and secrets services rather than self-hosted equivalents.
• Zero-downtime CI/CD. A GitHub Actions pipeline per service builds and scans the image, pushes to ECR and hands off to CodeDeploy blue/green: green tasks are health-checked, take traffic and bake under real load with instant rollback before blue is retired.
• Security and observability. Edge firewalling, threat detection and least-privilege IAM aligned to the AWS Startup Security Baseline; centralised metrics, logs and dashboards with alarms routed to the team; managed backups.

How the Platform Is Operated

After sign-off, SquareOps' 24×7 SRE team took operational ownership. A named account lead and a dedicated engineering group know the environment, so nothing needs re-briefing when something happens.

• Shift-based coverage every day of the year, with handover calls between shifts.
• Every alert becomes a prioritised ticket: P1 acknowledged within 15 minutes and targeted for resolution within 60, with root cause analysis for every P1.
• Production changes are scoped, costed and approved in writing before they are made, and monthly SLA and incident reports keep the client informed.

Year One Improvements

CloudFront in front of production, with no downtime. The production hostnames were moved behind Amazon CloudFront with the full WAF rule set at the edge, and the load balancer was locked to CloudFront's origin-facing IP ranges. Traffic was validated before DNS changed, and the cutover caused no downtime.

Every service moved to AWS Graviton. In July 2026 all seventeen services were rebuilt for ARM64. On staging, each Graviton service ran alongside its x86 twin so QA could verify it without touching live traffic; production was then cut over service by service with the client confirming each one. AWS prices Fargate on Graviton roughly 20% below x86, and the migration also lowered the estate's reported carbon footprint.

Non-production cost optimisation. A cost review found an unused OpenSearch domain, over-allocated services and missing lifecycle policies. Every finding was ticketed and executed, and the staging stack now shuts down outside working hours with a one-click restart for developers. Staging AWS spend fell by 18.5% within two months.

Resilience and hardening. Production services were raised to two running tasks across availability zones, shared database credentials were replaced with per-service credentials scoped to each service's own schema, and a disaster recovery assessment produced a phased roadmap now being scoped with the client.

Results

• Launched on schedule on a production-grade AWS platform, with no cloud operations hires on the client side.
• 99.9% uptime SLA with 24×7 coverage and 15-minute P1 acknowledgement.
• Zero-downtime releases for all seventeen services, and a zero-downtime CloudFront cutover.
• 100% of compute on Graviton, migrated without a single unplanned outage.
• 18.5% lower non-production AWS spend within two months of the cost review.

What the Client Says

"Having a dedicated 24×7×365 SRE team has made a real difference. They know our environment and don't need to be re-briefed every time something comes up. Beyond just keeping the lights on, they helped us stabilise our architecture and put proper security best practices in place. During production incidents they take ownership without us having to chase updates."