KubeCon India recap | Bitnami images | Open source maintainer fatigue?

KubeCon India marked my first KubeCon Keynote. It’s always a joy to meet the cloud-native community, and KubeCons have a special place in my heart as they’re truly the gathering ground for all cloud-native enthusiasts. This year at KubeCon India, I noticed three major themes:

🔍 Observability Our keynote was centered around observability. We showcased how to best use cloud-native tools together to observe applications including AI agents! The recording isn’t out yet, but keep an eye on the CNCF YouTube channel.

🤖 AI AI continues to be everywhere at conferences, and KubeCon is no exception. It makes perfect sense as most of us are infrastructure engineers, and in the end, all AI agents, apps, and training pipelines run on infrastructure. Building and scaling the right infrastructure for AI is one of the biggest priorities in 2025.

🛠️ Platform Engineering This has been a recurring theme across recent KubeCons. The reason is clear: every large enterprise is trying to empower their developers by building internal platforms using cloud-native tools. Platform engineering is becoming a cornerstone of modern enterprise strategy.

On a personal note, this KubeCon also marked my daughter’s 4th KubeCon and she’s only 3 years old! ❤️ I absolutely love the way she supports us. Both Saloni and I had sessions (hers was on LLMs on Kubernetes), and despite the challenges, we always find ways to do everything together, happily. Huge thanks to CNCF for the free childcare at KubeCon as it makes a real difference for tech parents.

The talks will be available soon on the CNCF YouTube channel so stay tuned!

Did you attend KubeCon India? What was your favourite moment? Do share in the comments!

Before KubeCon India, I had the privilege of being invited to judge the Infosys Global Hackathon in association with CNCF and the UN. I represented the CNCF side Round 1 included participants from 8 states, and then 32 selected teams competed in the final rounds at Infosys Hyderabad, just before KubeCon.

As a judge, I got to hear so many incredible ideas that people built, and in the end, I came back with a lot of learnings.

Bitnami pulling most free images is a reminder that ‘free’ in infra rarely lasts forever. Companies relying on it need to plan migrations, shift to upstream official images, fork and maintain what matters, or build internal images. Open source is powerful, but sustaining it has costs and depending blindly on free catalogs will always come with risks.

Maintainer fatigue is one of the biggest hidden risks in open source. As projects grow in adoption, the number of support requests, issues, and feature demands often skyrockets while the pool of active contributors remains small. This imbalance leads to burnout, where maintainers feel overworked, under appreciated, and sometimes resentful of the very community they helped build. External Secrets recent struggles are a reminder that “free” open source comes at the cost of human time and energy, and unless more organizations step up with real support for code, funding, or maintainers, the projects risk slowing down, moving to maintenance mode, or even being abandoned.

What are your thoughts on maintainer fatigue?

Kubesimplify as a media partner for KubeCon India

Kubesimplify was proud to be a media partner for KubeCon India, and we couldn’t have done it without the incredible support from the community that constantly pushes us to deliver the right and latest content. We recorded many interviews, which are now live across our socials along with some shorts.

• 3 things to learn before Kubernetes with Mumshad
• How to get selected at KubeCon feat Suman
• How to contribute to CNCF, women in cloud native with Bhavani
• Crossplane and k8sGB with Yury
• 3 reasons to attend Kubecon feat Atulpriya

Team Kubesimplify had a great time at KubeCon India, and we’re excited to continue the journey as official media partners for Container Days Hamburg, where we’ll be covering the event as well! Keep supporting us we’re on a mission to become the best tech media hub with an authentic touch.

Awesome Reads

• Supercharge your AI: GKE inference reference architecture, your blueprint for production-ready inference - Google Cloud has introduced the GKE inference reference architecture, a production-ready blueprint to simplify deploying AI/ML inference workloads on Kubernetes. Built on a secure, scalable base platform, it optimizes GPU/TPU usage, reduces cold starts, supports real-time, batch, and streaming inference, and abstracts complex model operations so teams can focus on innovation rather than infrastructure.
• Kubernetes 1.34: 10 new Alpha features - Kubernetes 1.34 introduces several exciting Alpha features, including smarter GPU allocation with Dynamic Resource Allocation (DRA), pod-level certificates, async scheduling APIs, improved YAML formatting, and in-place container restarts for AI/ML workloads. While not production-ready, these updates highlight Kubernetes’ focus on modern workloads, security, and operational resilience. What is the favourite feature you are looking forward to?
• Crossplane 2.0 - Crossplane 2.0 reimagines infrastructure and application management by making applications first-class citizens, introducing namespaced resources by default, and adding declarative day-two operations. With broader composition capabilities and managed resource filtering, it empowers platform teams to build full-stack self-service APIs while reducing complexity and improving multi-tenancy. What are your thoughts on crossplane in general and 2.0?
• PaC(Policy as Code) and AI - AI coding assistants speed up development but also risk introducing subtle bugs, insecure defaults, or inconsistent architecture, making Policy as Code (PaC) essential for governance. By automating policy evaluation, enforcement, and auditing, platform teams can let developers “stay in flow” while ensuring AI-generated code meets security, compliance, and resilience standards.
• Shipping an AI Agent that Lies to Production: Lessons Learned - Three Dots Labs shared lessons from building an AI Mentor for their learning platform, showing that while LLMs can help students debug and learn faster, they often hallucinate, fail unpredictably, and require heavy orchestration, testing, and cost management. The team found that success relied less on flashy AI tricks and more on solid engineering practices, careful evaluation, and setting limits to keep the system reliable and sustainable.

Awesome Resources/Repos

• Statistics and Data Analysis
• Gemini-cli as GitHub Actions
• How to Design Structured Database Systems Using SQL [Full Book]

Learn from X

• https://x.com/ai_christianson/status/1955715834904506461
• https://x.com/aakashg0/status/1955495290716492009
• https://x.com/Aurimas_Gr/status/1953417004083220563