<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Quince on Medium]]></title>
        <description><![CDATA[Stories by Quince on Medium]]></description>
        <link>https://medium.com/@onequince?source=rss-4f9080e2ae97------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ShAQu_YrQbcYjlX59xiQ0g.png</url>
            <title>Stories by Quince on Medium</title>
            <link>https://medium.com/@onequince?source=rss-4f9080e2ae97------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 09:38:06 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@onequince/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Short-Lived Application Testing Environment (SLATE) on EKS]]></title>
            <link>https://tech.quince.com/short-lived-application-testing-environment-slate-on-eks-eedfa18ffb15?source=rss-4f9080e2ae97------2</link>
            <guid isPermaLink="false">https://medium.com/p/eedfa18ffb15</guid>
            <category><![CDATA[quince]]></category>
            <category><![CDATA[infrastructure]]></category>
            <category><![CDATA[ephemeral-environment]]></category>
            <category><![CDATA[slate]]></category>
            <category><![CDATA[devops]]></category>
            <dc:creator><![CDATA[Quince]]></dc:creator>
            <pubDate>Fri, 13 Mar 2026 11:26:42 GMT</pubDate>
            <atom:updated>2026-03-13T11:30:29.104Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’ve ever shared a single DEV environment with multiple teams, you know the pain.<br>A new feature is deployed to the development environment, and unexpectedly, another team’s API tests start failing. Soon after, a different team deploys their branch, causing upstream integrations to break. What was once a smooth CI/CD pipeline has now turned into a daily cycle of trial and error. — <em>“who broke DEV this time?”</em></p><p>As our engineering teams grew, this chaos only amplified. Developers were constantly stepping on each other’s toes, waiting for their turn to test, or trying to debug issues that weren’t even theirs. Having one shared environment wasn’t cutting it anymore.</p><p>The obvious fix — giving each team their own dedicated DEV environment — looked great on paper, but in practice it is too expensive, too hard to maintain, and does not scale for a fast growing org like Quince. For larger teams with multiple developers working on different features simultaneously, the problem only intensified. Each developer often needed their own environment to test and take multiple features live in parallel, further increasing infrastructure costs and operational complexity. We needed a solution that could provide each pull request with an isolated sandbox environment, without the overhead of managing dozens of long-lived environments.</p><p>That’s when at Quince we decided to build <strong>SLATE </strong>— lightweight, short-lived ephemeral environments that can spin up on demand for every feature and tear down just as easily.</p><p>Developers can preview features, run integration tests, and validate releases in true production parity — without waiting on shared infra or impacting anyone else.</p><p>The result? Faster feedback loops, stable environments, and a developer experience that actually scales with the size of our teams.</p><h3>Challenges</h3><p>Before diving into implementation, it’s worth recapping the pain points that pushed us here:</p><ul><li><strong>Shared environment instability</strong> made feature testing slow and unreliable.</li><li><strong>Parallel development</strong> within the same team and across multiple teams forced sequential testing and coordination delays.</li><li><strong>Limited infra capacity</strong> meant we were always balancing stability vs. speed.</li></ul><p>We needed a model that allowed true parallelism, production parity, and self-serve provisioning — without blowing up costs or cluster complexity.</p><h3>Design Requirements</h3><p>Our design goals were shaped around developer autonomy and operational safety. We wanted something that just worked — predictable, governed and easy to use.</p><ul><li><strong>On-demand creation:</strong> PR-tied environments that can be spawned up manually, with strict TTL-based auto-cleanup. On-demand provisioning environments per PR gives developers fast, isolated feedback loops (tests, manual QA, demos) without waiting for shared env availability. TTL-based cleanup prevents resource leakage and keeps cloud costs and cluster complexity under control by ensuring environments are short-lived unless explicitly extended.</li><li><strong>Isolation:</strong> Each environment runs in its own namespace, argo-cd project, ingress, and shareable feature url.<br>Also, parity with production reduces “works on my machine” incidents by validating the feature against the same configuration, routing, and security constraints. Isolated namespaces and Argo CD projects limit blast radius and make rollback/teardown deterministic. A shareable URL enables reviewers and automated tests to interact with the exact running system.</li><li><strong>Dependency-aware fan-out:</strong> A dependency graph defines relationships between services. Developers can choose which dependencies get deployed and even pin versions.<br>Many features require multiple services to run together. A dependency graph automates the correct set of services, avoiding manual, error-prone wiring. Selective fan-out and version pinning let developers test with realistic dependency combinations (e.g., a stable backend with a feature branch frontend) and isolate failures to the intended change.</li><li><strong>Quota governance:</strong> Each team has a resource quota enforced before provisioning, based on estimated fan-out.<br>Quotas prevent a single team or runaway PRs from exhausting cluster capacity and impacting others. Evaluating quotas up front enforces fair resource allocation, enables capacity planning, and protects production/critical workloads from noisy neighbors.</li><li><strong>Cost awareness:</strong> All environments are tagged for cost attribution and expire aggressively by default.<br>Tagging enables chargeback/showback and helps teams understand the cost of CI/CD practices. Aggressive expiration limits idle spend and encourages responsible resource use; teams can request extensions when genuinely needed.</li><li><strong>Simple update loops:</strong> Developers can refresh or modify their environments directly from Bitbucket pipelines after code changes.<br>Seamless update paths keep environments synchronized with the latest code without manual intervention. Integrating with the existing CI (Bitbucket Pipelines) keeps the process familiar, reduces context switching, and enables deterministic, repeatable environment updates for iterative development and testing.</li></ul><p>These goals laid the foundation for a system that could scale seamlessly across teams and support parallel development workflows, while remaining fully automated and low-touch.</p><h3>Implementation</h3><p>Our ephemeral environment platform is built around a modular <strong>control plane</strong> and a <strong>convention-driven workflow</strong> that balances developer flexibility with operational safety.</p><h4>Control Plane</h4><p>At the heart of the system is a <strong>Flask-based microservice</strong> that acts as the control plane. It exposes APIs for all major operations:</p><ul><li>Creating new environments</li><li>Generating dependency YAMLs</li><li>Syncing and enforcing quotas</li><li>Cleaning up expired or idle resources</li></ul><p>This service is the single source of truth for environment orchestration, coordinating provisioning across Kubernetes, ArgoCD, and AWS infrastructure components.</p><h4>Dependency Modeling</h4><p>We introduced a canonical dependency.yaml to define service-to-service relationships.<br>It captures both direct and transitive dependencies, allowing the system to construct a complete dependency graph for any given service. This flexibility lets developers selectively enable or disable subsets of upstream services as needed based on the product feature that needs to be tested.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SeWMaxNguvw2k5m5d0683Q.png" /></figure><p>When a developer requests an environment, the control plane expands that graph into a normalized YAML. Developers can:</p><ul><li>Toggle which dependencies to include.</li><li>Pin builds or versions for specific services.</li><li>Based on the direct dependencies enabled, transitive automatically gets enabled.</li></ul><pre><br>main: service A<br>branch: feature/&lt;jira-ticket-id&gt;<br>pull-request: pr-88<br>dependencies:<br>- name: service B<br> branch: master<br> build: latest<br> enable: true<br>- name: service C<br> branch: feature/&lt;jira-ticket-id&gt;<br> build: build-22<br> enable: true<br>- name: service D<br> branch: master<br> build: latest<br> enable: false<br>transitive-dependencies:<br>- name: service E<br> branch: master<br> build: latest<br>- name: service F<br> branch: master<br> build: latest<br>.... and so on</pre><p>This gives each environment deterministic reproducibility with just enough flexibility to test real-world scenarios.</p><h4>Provisioning Workflow</h4><p>The provisioning workflow is fully automated and follows a predictable sequence:</p><ol><li><strong>Namespace creation:</strong> Each product feature gets its own Kubernetes namespace for isolation.</li><li><strong>Helm + ArgoCD integration:</strong> Helm values are customized per service, and ArgoCD Application specs are generated dynamically.</li><li><strong>Infrastructure bring-up:</strong> Ingress (ALB + TLS), unique application URL, Route53 DNS records, ESO, and a dedicated ArgoCD project are provisioned.</li><li><strong>Metadata persistence:</strong> Each environment’s details — status, TTL, owner, team, and “quota used” — are stored in DynamoDB.</li></ol><p>This approach ensures every ephemeral environment is complete, isolated, and fully auditable.</p><h4>Quota and Cleanup</h4><p>To prevent runaway provisioning, we built a <strong>quota and cleanup system</strong> on top of DynamoDB.</p><ul><li><strong>Preflight checks:</strong> Before provisioning, the control plane computes the estimated service count from the dependency graph and validates it against the team’s quota.<br>For example:<br>1. A <strong>standalone service</strong> with no downstream dependencies might have a quota of <strong>2</strong>, meaning only <strong>2 SLATE environments</strong> with one deployment per SLATE can exist at the same time for that service.<br>2. A service that depends on <strong>10 downstream services</strong> can have a higher effective quota say “20” — to accommodate the additional resources needed for all dependent services to run.</li><li>Developers can also selectively <strong>enable or disable specific downstream services</strong>, which adjusts the total number of SLATE environments provisioned. Depending on the combination of services enabled, a team might have anywhere between <strong>2 and 20 SLATE environments</strong> running in parallel at any given time for this example service.<br><strong>Key takeaway:</strong> The quota isn’t fixed per service; it scales based on the number of dependencies and which services are enabled, ensuring realistic testing environments without over-provisioning resources.</li><li><strong>Scheduled cleanup:</strong> A background job periodically scans for expired environments and cleans them up — deleting ArgoCD Applications, DNS entries, and Kubernetes namespaces, before marking them expired in DynamoDB.</li></ul><p>This blend of automation and governance keeps our clusters lean and predictable without human intervention.</p><h3>Developer Workflow</h3><p>The developer experience was intentionally kept very simple — no YAML wrangling, no kubectl gymnastics. Everything starts from a <strong>self</strong>-<strong>service</strong> <strong>developer portal</strong> that abstracts the complexity behind a clean UI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jqIGi_YHtOfWOR2MmXZjjQ.png" /><figcaption>Self service developer portal</figcaption></figure><h4>Dependency Information Generation</h4><p>Developers fill out a short form specifying:</p><ul><li>Repo name [Dynamic Dropdown]</li><li>Service name [Dynamic Dropdown]</li><li>Feature branch [Dynamic Dropdown]</li><li>PR number [Dynamic Dropdown]</li></ul><p>Based on the above provided information, dependency yaml is generated and post customizing it with the available specification, it is ready to feed in the next step.</p><h4>Environment Provisioner</h4><p>Developers fill out a short form specifying:</p><ul><li>Team Name [Dropdown]</li><li>Dependency File Upload [File Input]</li><li>Time to Live [TTL] [Dropdown]</li></ul><p>Once submitted, the portal calls the control plane microservice, which:</p><ol><li>Validates the requested input.</li><li>Checks for available quota for the team.</li><li>Generates Kubernetes Manifests for Namespace, ArgoCD Project and ArgoCD Applications.</li><li>Applies the generated Kubernetes Manifest, in the order of generation.</li><li>Configures ingress, DNS, and relevant dependencies.</li><li>Returns links to the ArgoCD project and a host url.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/782/1*gU3pFMkXcNIV28VXa89P8g.png" /><figcaption>SLATE user flow</figcaption></figure><p>Within a few minutes (&lt; 10mins), developers get a fully functional environment — isolated, production-like, and ready to test.</p><h4>Updating Environments</h4><p>After pushing code changes, developers simply trigger a <strong>Bitbucket pipeline, </strong>with few custom fields, to re-sync their environment.<br>This updates both the main service and any impacted dependencies automatically, ensuring the environment always reflects the latest commit state.</p><h4>Teardown and Expiry</h4><p>Every environment is ephemeral by design — it only lives as long as it needs to.</p><ul><li><strong>Automatic cleanup:</strong> Once the TTL expires, a scheduled job decommissions all resources and updates the record in DynamoDB.</li><li><strong>Manual teardown:</strong> Developers can also delete environments on demand directly from the portal.</li></ul><p>This self-regulating lifecycle keeps clusters clean, costs predictable, and the developer experience frictionless.</p><h3>Conclusion</h3><p>Building SLATE fundamentally changed how we develop and ship software. What used to be a bottleneck is now a superpower.</p><ul><li><strong>Decreased lead time for changes:</strong> Teams ship faster by eliminating shared-env contention.</li><li><strong>Lower change failure rate:</strong> Production-parity testing catches integration issues earlier.</li><li><strong>Independent, parallelized testing:</strong> Every PR gets its own isolated sandbox.</li><li><strong>Governed, cost-aware scale:</strong> Quotas, TTLs, and tagging keep usage predictable without sacrificing autonomy.</li></ul><p>By investing in SLATE , we didn’t just improve our CI/CD pipeline — we made our development process scalable, reliable, and fast enough to keep up with our teams.</p><p><strong>Roadmap</strong></p><p>Our next goals focus on extending SLATE beyond Kubernetes resources to fully provisioned environments for each PR:</p><ol><li><strong>Ephemeral AWS resources:</strong> Automatically create and tear down cloud resources (DB, S3, REDIS etc.) per PR to achieve end-to-end testing parity using Crossplane+Terraform</li><li><strong>Fresh data seeding:</strong> Provide each feature with an isolated set of stateful data, ensuring realistic test scenarios without interfering with other environments.</li></ol><p>Author: <a href="https://medium.com/u/84cd2bbb91ed">Vatsal Dhar</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eedfa18ffb15" width="1" height="1" alt=""><hr><p><a href="https://tech.quince.com/short-lived-application-testing-environment-slate-on-eks-eedfa18ffb15">Short-Lived Application Testing Environment (SLATE) on EKS</a> was originally published in <a href="https://tech.quince.com/">Quince Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Meet Quince]]></title>
            <link>https://tech.quince.com/meet-quince-9a4b92466f5e?source=rss-4f9080e2ae97------2</link>
            <guid isPermaLink="false">https://medium.com/p/9a4b92466f5e</guid>
            <category><![CDATA[quince]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[e-commerce-business]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Quince]]></dc:creator>
            <pubDate>Mon, 08 Nov 2021 08:46:45 GMT</pubDate>
            <atom:updated>2021-11-15T19:05:00.314Z</atom:updated>
            <content:encoded><![CDATA[<p>At Quince, we are challenging the notion that nice things are expensive. We make quality-first essentials easy and attainable — from apparel and fine jewelry, to home decor and kitchenware. We’re powered by our Manufacturer To Consumer (M2C) approach and work with the world’s leading suppliers to bring the best possible products straight to your door, without the massive markups associated with traditional brands and retailers. We believe that high-quality, well-made products should be accessible to everyone.</p><p>Quince makes everyday indulgences attainable. We offer a range of clothing styles and accessories from dresses, loungewear, and everyday tees, to leather goods and fine jewelry, as well as home essentials such as sheets and bedding, robes and spa towels, kitchenware, rugs, and more. Only the best quality items from top suppliers worldwide can make their way to our customers.</p><p><strong><em>Our Model</em></strong></p><p>On the heels of the Direct to Consumer evolution, Quince represents the future of retail with our M2C model and proprietary software.</p><p>We analyzed which categories have the highest markups caused by complicated and overcrowded supply chains, and revised every part of the traditional retailer’s playbook by cutting out the middleman (and ourselves), which is why specialist factories are the beating heart of Quince. We search around the globe from India and Italy to Turkey, looking for the best of the best suppliers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rFiFOpKeYZzAWsIa" /></figure><p>We have built quite some software from the ground up and leveraged our expert teams in data insights and manufacturing to collaborate on product development to offer incredible products we know our customers will love. Our advanced technology and deep integration with our factory partners allows us to produce the right quantities, eliminating financial and environmental waste.</p><p>We’re committed to investing in deep, cost-optimized relationships with our manufacturing partners to provide immense value to both the factories and our end-customer. Through this, we are able to curate the highest quality products made by the finest factories at the lowest possible price, which simplifies the shopping experience, specifically compared to other marketplaces.</p><p><strong><em>The Quince Standard</em></strong></p><p>At Quince, we believe sustainability should be a requirement, not a luxury. We work exclusively with factories that reflect the same high environmental and ethical values that we hold.</p><p>We look to partner with mid-sized, family-run operations that have employed the same workforce for many years, but we also work with larger operations that employ rigorous ethical standards. This means paying workers fairly and doing their best to produce in a way that reduces damage to the environment.</p><p>We also reduce our carbon footprint by shipping directly from our factories to you. Importantly, our core mission of offering high-quality essentials that customers actually want to buy means that our clothing and products will be worn or used for a long time, so they won’t end up in landfills.</p><h3><strong><em>Engineering At Quince</em></strong></h3><p>Quince is not just a brand or just an e-commerce company. Behind the scenes, we’re a full-fledged supply chain and logistics platform. This expands the horizon of engineering at Quince from e-commerce to a plethora of tools and technology for supply chain and logistics. Our engineering team primarily operates in 5 verticals right now:</p><p><strong><em>E-commerce vertical</em></strong>: focuses on building the storefront and the e-commerce platform. This team handles the growing demand by building scalable solutions keeping customer experience at the forefront.</p><p><strong><em>Growth vertical</em></strong>: focuses on optimizing customer acquisition cost and repeat rate. Centered around data, the goal is to optimize each and every channel and touch-point with customers - showing useful and interesting content to the right audience on the right channel at the right time.</p><p><strong><em>Logistics vertical</em></strong>: when the growth team gets orders, the logistics tech focuses on delivering shipments from 8 different countries to the customer’s doorstep with everything automated in between — customs, country-specific paperwork &amp; taxes, shipment tracking, and exception workflows.</p><p><strong><em>Supply-chain vertical</em></strong>: this vertical caters to the complex world of supply-chain with the aim to simplify it, starting from finding the next Quince product to getting it ready for sell—product market research, merchandizing, product lifecycle management, inventory planning, purchase-order management, and warehousing.</p><p><strong><em>Finance vertical</em></strong>: making sure that the unit economics makes sense. The tooling starts from understanding the various cost components, doing reconciliations, calculating margins, and analyzing pricing to building workflows for invoicing and cross-border payments.</p><p>As of the date of writing this blog, Quince engineering is 17 people strong. Data architecture, scale architecture, and infrastructure are all shared concerns amongst all these groups. Data science seamlessly aligns with all the verticals. As the teams grow, these verticals will be broken down into finer domains with much more closely knit KPIs.</p><p>All these engineering verticals are solving many interesting problems. Something we would talk more about in the next set of blogs along with our technology architecture, design choices, our data-driven methodology, and the tech stack, that helped us move fast in our zero-to-one phase effectively and efficiently.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9a4b92466f5e" width="1" height="1" alt=""><hr><p><a href="https://tech.quince.com/meet-quince-9a4b92466f5e">Meet Quince</a> was originally published in <a href="https://tech.quince.com/">Quince Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>