KubeCon + CloudNativeCon India 2026 brought the cloud native community to Mumbai on June 18-19. For LitmusChaos, this was not just another conference. It was one of our most significant events to date. We walked away with a keynote slot, a packed project booth, and conversations that will shape how we think about growing this community in India and beyond.
Flipkart and LitmusChaos: The CNCF End User Case Study contest and keynote
The highlight of the event for us came before a single session even started. Flipkart was selected as the winner of the CNCF End User Case Study Contest India, with a story that featured LitmustChaos. This honor earned us a keynote slot alongside Flipkart on the main stage at KubeCon India 2026.
This recognition is a reflection of the real-world impact LitmusChaos is having at scale. It is not just about the project itself. It is about the organizations that have adopted it, built on top of it, and contributed their learnings back to the community. Flipkart’s story is exactly that.
Keynote: From afterthought to practice
From Afterthought to Practice: How Flipkart Built a Multi-tenant Chaos Platform on LitmusChaos Aditya Sridasyam, Software Development Engineer, Flipkart, and Uma Mukkara, Head of Resilience Testing
Flipkart runs hundreds of tightly coupled microservices that need to hold up under the kind of traffic that Big Billion Days and festive sales generate. For a long time, resilience was something the team dealt with after an outage, not before. To change that, Flipkart’s Central Reliability Engineering team built a centralized chaos platform on top of LitmusChaos.
What made this talk stand out was the specificity. Aditya walked through four concrete customizations Flipkart made to run LitmusChaos at their scale: a hybrid multi-tenancy architecture that sits between cluster-wide and namespace-wide installs, a DaemonSet-based high-availability model for chaos injection, a Script Runner fault that enables dynamic target selection and context chaining, and a hybrid VM chaos extension for workloads that are not running on Kubernetes. These are the kinds of production-hardened decisions that come from running chaos engineering at the scale Flipkart operates at.
Uma Mukkara provided context on how contributions like these flow back upstream and benefit the broader LitmusChaos community.
For a full look at Flipkart’s reference architecture, platform design, and the measurable impact numbers, read the CNCF Flipkart case study.
Project pavilion conversations
Our project booth at the pavilion saw between a hundred and two hundred visitors. The conversations ranged from first-time introductions to chaos engineering to deep technical discussions on architecture and integrations. Below is a summary of what came up most at the booth.
Resilience Testing and Reliability in the Age of AI: AI inference workloads are fragile in ways traditional services are not, and a lot of the conversations at the booth were about how chaos engineering fits into that new reality.
ChaosHub: The Default Fault Library: We walked many people through ChaosHub, LitmusChaos’s ready-to-use fault library covering Kubernetes, Linux, AWS, GCP, and more. For most, it was the first time they realized how much is available out of the box.
Automating Chaos in CI/CD: A common ask was how to shift chaos left. We covered integrating LitmusChaos into pipelines using LitmusCTL, the SDKs, and Terraform without requiring teams to overhaul their existing delivery process.
Tool Comparisons and Ecosystem Positioning: We had several side-by-side comparisons with other chaos engineering tools, focusing on where LitmusChaos fits and why the CNCF-backed, community-driven model matters for long-term adoption.
Custom Chaos and Best Practices: For teams with infrastructure patterns not covered by the default library, we discussed how to build custom experiments and what a mature chaos engineering practice looks like at different stages.
LitmusChaos MCP: We introduced the LitmusChaos MCP, a Model Context Protocol integration that lets engineers interact with and learn chaos engineering through natural language, lowering the barrier to entry for teams just getting started.
New Adopters: We shared that Canonical and Intertech have recently joined as official adopters. Flipkart’s story on the keynote stage the same day added strong validation of the project’s production-readiness to that conversation.
Alongside the technical conversations, we ran a Chaos Bird game at the booth, a leaderboard-based take on Flappy Bird where the top scorers won LitmusChaos swag. We also ran a separate random giveaway throughout both days. The engagement was high, and the leaderboard kept people coming back.
A word on what this event meant
The KubeCon + CloudNativeCon India audience is largely practitioners building and running systems in India, working with constraints and at scale. The interest in LitmusChaos here was genuine and grounded. The questions were specific. That is exactly the kind of community engagement that makes a project better.
Sharing Flipkart’s story for the winning CNCF End User Case Study Contest is a milestone for the project and for the Indian cloud native community. We are grateful to Aditya and the Flipkart team for sharing their work and to Uma for anchoring the session. We are also grateful to CNCF and the event organizers for creating the space for this kind of recognition.
Stay connected with LitmusChaos
If KubeCon + CloudNativeCon India was your first introduction to LitmusChaos, here are some ways to go deeper:
- Visit the LitmusChaos website for architecture, use cases, and getting-started guides.
- Join the #litmus channel on Kubernetes Slack to ask questions, share experiments, or talk to the maintainers directly.
- Explore ChaosHub to browse the ready-to-use fault library.
- Check out the Contributing Guide if you want to contribute issues, suggestions, or PRs.
- Subscribe to the LitmusChaos YouTube channel for community meetings, demos, and tutorials.
- Follow @LitmusChaos on X and @litmuschaos on LinkedIn for project updates.
- Write about your own experience with Litmus on DEV.to and tag it with #litmuschaos.
Thank you to everyone who stopped by the booth, attended the keynote, and made KubeCon + CloudNativeCon India 2026 a memorable event for the LitmusChaos community.
Facts Only
* KubeCon + CloudNativeCon India 2026 took place June 18-19 in Mumbai.
* Flipkart won the CNCF End User Case Study Contest India.
* Aditya Sridasyam and Uma Mukkara of Flipkart delivered a keynote presentation.
* Flipkart implemented a centralized chaos platform using LitmusChaos.
* Flipkart's technical customizations include a hybrid multi-tenancy architecture, a DaemonSet-based high-availability model, a Script Runner fault, and a hybrid VM chaos extension.
* Between 100 and 200 people visited the LitmusChaos project booth.
* LitmusChaos introduced a Model Context Protocol (MCP) integration for natural language interaction.
* Canonical and Intertech joined as official adopters of LitmusChaos.
* The LitmusChaos project includes a fault library called ChaosHub.
* Event activities included a "Chaos Bird" leaderboard game and random giveaways.
Executive Summary
LitmusChaos utilized KubeCon + CloudNativeCon India 2026 to demonstrate its production viability through a high-profile partnership with Flipkart. By winning the CNCF End User Case Study Contest, the project secured a keynote stage to showcase how Flipkart transitioned resilience testing from a reactive "afterthought" to a proactive practice. The presentation focused on specific architectural adaptations—such as hybrid multi-tenancy and VM extensions—required to manage chaos engineering at the scale of major Indian retail events.
Beyond the keynote, the project focused on ecosystem expansion and accessibility. Engagement at the project pavilion centered on integrating chaos engineering into AI workloads and CI/CD pipelines, while the introduction of an MCP integration aimed to lower technical barriers for new users. The addition of Canonical and Intertech as adopters suggests a growing trajectory for the tool. While the event's success is framed through high booth traffic and keynote recognition, the primary value lies in the public validation of LitmusChaos's ability to handle extreme enterprise scale.
Full Take
The strongest version of this narrative is that an open-source resilience tool has reached a critical inflection point, moving from experimental utility to enterprise-grade infrastructure capable of supporting some of the world's highest-traffic retail environments.
This is a classic vendor-driven success story. The narrative relies on a "Proof of Scale" pattern: by tethering the project's identity to Flipkart's massive traffic, the implied conclusion is that if it works for the "Big Billion Days," it will work for any user. The integration of a "Chaos Bird" game and swag giveaways alongside deep technical discourse creates a bridge between community-building and professional validation. This is a calculated effort to humanize a highly complex technical product while simultaneously asserting its dominance through the "Authority Game"—using a CNCF award and a massive corporate logo to signal readiness.
Patterns detected: ARC-0023 Authority Game
The underlying paradigm is the "shift-left" movement in DevOps, assuming that failure is inevitable and therefore must be engineered into the development lifecycle. This benefits the tool creators and the reliability engineers, but the second-order consequence is a permanent increase in system complexity; the "cure" for fragility is the introduction of an additional layer of sophisticated chaos-injection software.
If this were a coordinated influence campaign, the playbook would be "Validation by Association": selecting a high-visibility regional champion (Flipkart), securing a sanctioned institutional stamp of approval (CNCF), and flooding the community with "low-friction" entry points (MCP/Natural Language). The actual content aligns with this pattern, as it is essentially a marketing-adjacent summary of a corporate win.
How does the reliance on a single "mega-user" case study mask potential failures in smaller or differently structured environments? What specific metrics—beyond "measurable impact numbers" mentioned vaguely—would prove the platform's efficacy? Who bears the cost if the "chaos" engineered into a system causes an unforeseen cascading failure during a peak event?
Sentinel — Human
The text reads like a professional summary of a conference presentation, successfully blending factual reporting with promotional enthusiasm, indicating a likely human origin focused on community engagement.