In the January 2026 edition: in-place Pod vertical scaling goes GA in Kubernetes v1.35, why QA is shifting left into platform engineering, and how Netflix balances efficiency with reliability. The Optimal is the Akamas newsletter where DevRel Engineer Graziano Casto breaks down Kubernetes performance, reliability, and cost optimization.
Hey there, there is a new captain at the Helm!
I’m Graziano, and I’m taking over The Optimal. If we haven’t met yet, I’m the DevRel here.
Why the change?
Because I’m a firm believer that the best technical insights come from real conversations. I want to make sure every email you get from us feels like a helpful chat between peers, sharing what actually works in the trenches.
From now on, we’re bringing a fresh perspective to this space. We’ll dive into what’s actually happening in the world of reliability, performance, and maybe share a laugh about why cloud bills still look like phone numbers from the 90s. Expect plenty of practical “hot takes”, some well-placed memes, and absolutely zero fluff.
Before we dive into the logs, a quick health check: is your current CPU usage lower than your stress levels?
If so, you’re doing great! Let’s scale.
The Latest Replica
Kubernetes: No more restarts for resizing?
With Kubernetes v1.35, “In-place Pod Vertical Scaling” is finally GA. For those of us obsessed with optimization, this is a game-changer. Being able to adjust resources without killing the pod means we can finally aim for true dynamic efficiency without the nightmare of cold starts. It’s the first real step toward an ecosystem that reacts in real-time.
QA is moving to the Platform
The organizational landscape is shifting, and this article is a loud wake-up call. Responsibility for performance and reliability is performing a hard shift-left directly into the platform. QA is evolving from a “gatekeeper” at the end of the cycle into the team that builds the automated guardrails to make systems resilient by design. Less manual testing, more engineering.
How Netflix balances the “Unbalanced”: Efficiency vs. Reliability
If you haven’t watched this re:Invent session yet, clear your schedule. Netflix explains how they manage a global fleet where a single minute of downtime costs ~$200k. My biggest takeaway? They don’t aim for 50% utilization across the board. Instead, they intentionally run critical “tier 0” services cooler (30%) to buy safety, while pushing batch workloads to 80%+. It’s a masterclass in risk-adjusted value, precisely the kind of “systems thinking” we advocate for here at Akamas.
Commits From The Lab
I’m thrilled to announce that our Free Trial is finally available for everyone! You can now get your hands on Akamas Insights with all features unlocked for two weeks. The goal? To give you concrete, actionable recommendations on how to improve both the costs and reliability of your clusters. No more guessing games, just data-driven optimization.
Give it a spin here!
If you run into any issues, have questions, or just want to share some feedback, I’m always available. Don’t hesitate to reach out. I’m here to help you get the most out of it.
Catch Us
Want to meet the Akamas team in person? We’re regularly at conferences and meetups across the Kubernetes and Java ecosystems. See where we’ll be next on our Events page, and follow me on LinkedIn for the talks and sessions I’ll be at.
That’s it for this first run. Subscribe using the form, and you’ll get every edition in your inbox – then just hit reply with a topic you want me to cover or a performance horror story to share. I read every message.
Stay optimized,
Graziano Casto, DevRel @ Akamas
