Picture this: You’re standing at a packed platinum sponsor booth, answering your 50th question about Kubernetes cost optimization while simultaneously coordinating a mega afterparty, all while your voice is slowly giving up. That was me at KubeCon India 2025, and honestly? I’d do it all over again in a heartbeat.
The journey before the journey
My KubeCon adventure didn’t start when I walked into the HICC Hyderabad venue. It began days earlier with reunions that reminded me why our cloud native community is special.
Landing in Hyderabad, I found myself speed boating at Hussain Sagar Lake with an incredible crew – Hrittik Roy, Shivay Lamba, Animesh Pathak, and the dynamic duo Prateek and Juhi Singh. Between the splashing water and laughter, we were already discussing Kubernetes patterns and DevOps challenges. Only in India would you find engineers debating service mesh architectures while racing across a lake! The evening ended with authentic Hyderabadi biryani at Shadab – because great tech conversations deserve great food.
But the real warm-up? Mentoring at the CNCF UN Global Hackathon at Infosys. Alongside legends like Saiyam Pathak, AtulPriya Sharma, and Raghu Reddy, we guided teams through 48 hours of innovation. Watching fresh minds tackle real-world problems with cloud-native solutions reminded me why developer advocacy isn’t just a job…it’s a mission.
Out of 33 outstanding teams, the TechForGood 2025 Finals crowned Agriguru the first-place winner for its innovation, impact, and open-source excellence in Sustainability. The project was created by Samyuktha S, Manish M, Madhan S, Jyoti Prakash Behera, and Deepika Samad and exemplifies the power of technology for social good.
Cast AI takes center stage
When KubeCon officially kicked off, Cast AI’s booth became the epicenter of cost optimization conversations. Our booth wasn’t just busy – it was packed. Engineers, CTOs, and DevOps teams lined up to understand how they could slash their Kubernetes costs without compromising performance.
The APA Hero movement: next-gen engineers leading the charge
One of the most exciting moments was recognizing our APA Heroes – the pioneers who’ve completed our comprehensive certification program covering Monitoring, Node Autoscaling, and Kubernetes Security. These aren’t just certified professionals; they’re the vanguard of a new engineering paradigm.
What makes an APA Hero special?
They represent the next generation of engineers who understand that manual tuning in today’s cloud-native environments is like using a compass to navigate space. They’ve mastered Application Performance Automation – the emerging standard for automating performance, cost, and security across Kubernetes environments.
The highlight?
Watching Abhishek Veeramalla personally hand over exclusive APA Hero swag to our certified champions. These engineers received exclusive swag and, most importantly, recognition as leaders in the automation revolution. They were also given exclusive invitations to our invite-only KubeCon afterparty – because excellence deserves celebration.
The numbers speak for themselves: hundreds of meaningful conversations, dozens of live demos, and countless “aha!” moments when teams realized they were overspending on their clusters by 50-70%, Almost 100+ engineers signed up to become APA Heroes on the spot, Every conversation reinforced a simple truth – in 2025, optimizing cloud costs isn’t optional; it’s survival.
The ShareChat story that stopped the crowd
The highlight is our demo session, “Maximizing Efficiency of Reservations & Commitments in Kubernetes,” which features ShareChat’s journey.
Standing room only. That’s what we had when Gaurang Singh (our Field CTO APAC) and Abhiroop Soni from ShareChat took the stage. As one of India’s largest Google Cloud users, ShareChat’s story resonated deeply with the audience.
Here’s what made jaws drop:
- The challenge: Managing Committed Use Discounts (CUDs) and Reservations across thousands of nodes while serving millions of users
- The reality check: Traditional approaches left 30-40% of reservations underutilized
- The Cast AI solution: Automated optimization that intelligently manages workload placement to maximize CUD utilization
- The results: ShareChat achieved significant cost reductions while actually improving performance
The technical depth was just right; we showed real Kubernetes manifests, actual cost dashboards, and live optimization in action. The audience wasn’t just listening; they were taking photos of slides, asking detailed follow-ups, and many signed up for trials right there.
Technical gems: talks that caught my attention
While booth duty kept me busy, I managed to catch some incredible sessions that showcased the cutting-edge of cloud native technology:
“Freeze, Investigate, Recover: Unlocking Forensic Container Checkpointing With CRI-U” by Suman Chakraborty (Platform9 Systems) and Neel Shah (Middleware) was a standout. The concept of using CRIU for container-level checkpointing opens fascinating possibilities for security forensics and failure recovery. Instead of losing valuable runtime data during pod restarts, this approach enables snapshots of complete application states – a game-changer for debugging and resilience.
“Fine-Grained Authorization: The Missing Piece in Agentic AI Security” by Shivay Lamba (Couchbase) and Ashish Jha (Okta) addressed a critical challenge in AI security. Their exploration of implementing permission models for RAG and agentic AI systems using OpenFGA and LangChain demonstrated how enterprises can prevent data leakage while scaling to billions of access decisions.
The keynote “From Outage To Observability: Lessons From a Kubernetes Meltdown” by Saiyam Pathak (LoftLabs) and Arnab Chatterjee (Nomura) resonated deeply with anyone who’s lived through production incidents. Their journey from basic Prometheus/ELK setups to distributed observability with proper multi-tenancy and fine-grained autoscaling provided valuable lessons for building resilient systems.
The human side of tech
What made this KubeCon special wasn’t just the technology; it was the people. I met Subham Londhe and discovered he’s from Bangalore (future collaboration loading… 🚀). I connected with Sagar Utekar and planned something exciting for Pune. These aren’t just networking moments; they’re the seeds of future innovations.
Our Cast AI afterparty on Day 1 wasn’t just an event – it was a celebration of the community. Seeing faces like Nasirullah Chowdhury (the Cloud Champ himself!), Abhishek Veeramalla, Soham and Suman Chakraborty, and Josh Berkus all in one room, sharing stories and laughs – this is what makes our ecosystem thrive.
Key takeaways you can act on today
After hundreds of conversations, here’s what the Indian Kubernetes community needs to know:
1. Cost optimization is not cost-cutting
Teams often fear that reducing costs means reducing capabilities. ShareChat proved otherwise – they actually improved performance while cutting costs. It’s about being smart with resources, not starving your applications.
2. Reservations are gold – if you use them right
Most organizations purchase Reserved Instances or CUDs but fail to utilize them effectively. Automated placement and workload optimization can increase utilization from 60% to 95%+.
3. The multi-cloud reality is here
Almost every enterprise booth visitor was running workloads across multiple clouds. The challenge isn’t choosing one cloud – it’s optimizing across all of them.
4. Automation beats manual optimization every time
The days of spreadsheet-based cost management are over. The scale and complexity of modern Kubernetes deployments demand intelligent automation.
5. Community is your competitive advantage
The connections made at KubeCon – from hackathon mentoring to booth conversations – these relationships drive innovation. The Indian cloud native community is not just growing; it’s leading.
What’s next?
As I write this, taking a well-deserved break after an incredible week, I’m already planning how to implement everything learned and discussed. The conversations at KubeCon weren’t just talk; they were commitments to build better, more efficient systems.
For those who couldn’t make it to KubeCon, here’s your action plan:
- Connect your cluster to Cast AI – Start with our free tier. Connect any cluster and see your potential savings in minutes. No credit card, no commitment, just insights. Get started here →
- Level up your skills – Complete our APA Hero Certification. It’s not just about Cast AI – it’s about mastering Kubernetes cost optimization principles that apply everywhere. Start learning →
- Join the conversation – Our Slack community is where the real discussions happen. From troubleshooting to best practices, you’ll find answers and make connections. Join our Slack →
- Book a personalized demo – If you’re managing significant Kubernetes workloads, let’s talk. Our team can show you exactly how much you could save. Book a demo →
Final thoughts
KubeCon India 2025 wasn’t just an event; it was a statement. The Indian tech community is not just consuming cloud native technologies; we’re shaping them. From ShareChat’s massive scale optimizations to startups just beginning their Kubernetes journey, everyone is pushing boundaries.
The cost optimization conversation has evolved from “nice to have” to “must have.” In a world where every rupee saved on infrastructure can be invested in innovation, can you afford not to optimize?
As I prepare for upcoming collaborations in Bangalore and Pune, and continue building the CNCF Kolkata community, one thing is clear: the best part of KubeCon wasn’t the technology showcased – it was seeing our community’s hunger to build better, more efficient systems.
Thank you to everyone who stopped by our booth, attended our session, or joined our afterparty. To those who missed it, see you at the next one. The cloud native revolution in India is just getting started, and trust me, you want to be part of it.
Until next time, keep optimizing, keep building, and remember – your Kubernetes clusters probably cost too much. Let’s fix that together. 🚀
Kubernetes cost optimization
Monitor organization-wide and cluster-level resource spending. Automate resource allocation and scale instantly with zero downtime.
