Tech Blogs on Ishan Jain

From Huffman to OpenObserve: Why Your 10GB Log File Fits in 3GB

Thu, 30 Jul 2026 00:00:00 +0000

Every SRE has run gzip on a log file and watched 10GB collapse into 3GB. It feels like magic. It is not. It is a handful of very old, very simple ideas about redundancy, and once you understand them, the storage numbers behind modern observability tools stop looking like marketing and start looking like arithmetic.

Let me walk from first principles up to how a tool like OpenObserve turns those principles into a 140x cost gap.

One Valkey, Multiple Tenants: Per-Database Isolation on Kubernetes

Wed, 27 May 2026 00:00:00 +0000

How Valkey 9.1.0’s per-database ACL selector lets you run multiple microservices off one shared cache instance, with hard access boundaries between them.

Stop Asking for a DevOps Roadmap. Read This Instead

Tue, 19 May 2026 10:00:00 +0530

Stop Asking for a DevOps Roadmap. Read This Instead

People who reach out to me and ask “give us a roadmap of becoming a DevOps engineer” or “what was your roadmap” - I go blank. What was my roadmap? I end up saying: learn Docker, K8s, Terraform, or AWS.

So I sat down and wrote my thoughts. If I needed to start again with no memory of being a DevOps engineer and read this blog, I should be able to get all the skills required to be good enough.

The Curious Case of Docker Hardened Images

Sun, 10 May 2026 00:00:00 +0000

An investigative look at Docker Hardened Images: what changed in December 2025, how DHI compares to distroless and Wolfi, and what the operational tax really looks like.

The Curious Case of Request and Limit on AWS EKS Fargate

Sun, 02 Nov 2025 14:53:00 +0530

If you’re running containers on AWS EKS with Fargate, you might have encountered a surprising reality about how billing works. Let me share a scenario that trips up many engineers:

The Scenario

You have a pod with the following resource specifications:

resources:
  requests:
    cpu: 100m
    memory: 500Mi
  limits:
    cpu: 4000m
    memory: 8000Mi

At first glance, you might think: “Great! My pod will use minimal resources (100m CPU, 500Mi memory) but can burst up to 4 vCPUs and 8GB when needed.”

The Most Practical Way to Get Started on AWS: Build a Producation Grade Cloud Setup

Wed, 17 Sep 2025 10:00:00 +0530

The Most Practical Way to Get Started on AWS: Build a Producation Grade Cloud Setup

When people say ‘learn AWS,’ it’s easy to get lost in a sea of services. In reality, most real-world architectures start with a VPC, subnets, instances, storage, and permissions.
In this guide, we’ll build a secure, production-style setup step-by-step — and we’ll do it four ways:

AWS Console
AWS CLI
Bash Automation Scripts
Terraform

What You Will Gain After Doing This

Hands-on experience with core AWS services
Learn how to automate (CLI, scripts, Terraform)
Cover security best practices (bastion, private DB, IAM roles, S3 VPC endpoints)

Tech Blogs on Ishan Jain

From Huffman to OpenObserve: Why Your 10GB Log File Fits in 3GB

One Valkey, Multiple Tenants: Per-Database Isolation on Kubernetes

Stop Asking for a DevOps Roadmap. Read This Instead

Stop Asking for a DevOps Roadmap. Read This Instead

The Curious Case of Docker Hardened Images

The Curious Case of Request and Limit on AWS EKS Fargate

The Scenario

The Most Practical Way to Get Started on AWS: Build a Producation Grade Cloud Setup

The Most Practical Way to Get Started on AWS: Build a Producation Grade Cloud Setup

What You Will Gain After Doing This

AWS Architecture Diagram