Services

Resources

Company

Our Work

Blog

Browse by topic.

Case studies, blog posts and talks from One2N engineers, filed by what they cover.

Reliability and operations

SRE, observability, incidents and performance.

Disaster Recovery

Incident Management

Monitoring

Observability

Open-Telemetry

OpenSRE

Performance Optimization

Resiliency

Scalability

SRE

SRE-Math

Cloud and platforms

Kubernetes, AWS, infrastructure as code, CI/CD and cost.

AWS

CI/CD

Cloud Migration

Cost Optimization

Data Migration

DevOps

DevSecOps

GIT

Infrastructure

Infrastructure As Code

Istio

Kubernetes

Networking

Platform Engineering

Security

Virtualization

Data and AI

Pipelines, streaming, databases and AI in production.

AI

Analytics

CDC

Clickhouse

Data Engineering

Debezium

Kafka

MLOps

PostgreSQL

SpringKafka

Software engineering

Architecture, languages and how we build.

Architecture

Backend Engineering

Best Practices

Distributed Systems

Go

How To

Lua

Open-Source

Overview

Industries

Work filed by the sector it was built for.

Online Gaming

Teams and careers

Leadership, mentoring and life at One2N.

Bootcamp

Life@One2N

Mentorship

Team Topologies

Technical Leadership

Case studies

Case studies

AI

AWS

Cloud Migration

Cost Optimization

Data Engineering

Data Migration

Go

Kubernetes

Lua

Monitoring

Observability

Online Gaming

Open-Telemetry

Performance Optimization

Platform Engineering

Resiliency

Scalability

SRE

Read Case Study

Kong API Gateway for Gaming and Sports Betting Integrations
How we used Kong Gateway and Lua to decouple legacy wallet systems and accelerate global expansion for a Tier‑1 gaming enterprise.

Read Case Study

Modernising Observability for a Billion Dollar Revenue Gaming Platform
A gaming and betting operator ran 200+ Java services across 30,000+ VMs on AWS and on-premises, processing 180M spins daily and 600K RPS, but needed unified observability across fragmented monitoring tools.

Read Case Study

Migrating from AWS Glue to Self-Hosted ClickHouse: 70% Cost Reduction for a Fintech Client.
How One2N cut a fintech client's analytics spend from $13K to $4K/month by replacing AWS Glue, DMS, and Athena with a self-hosted ClickHouse stack.

Read Case Study

How we reduced Kafka message latency by 90% and increased throughput 4x for a $300M SaaS platform.
How One2N reduced Kafka message latency by 90% and increased throughput 4x for a $300M SaaS platform, while cutting AWS infrastructure costs by 40%.

Read Case Study

Journey of moving VM-based workloads to Kubernetes.
Flip is a fintech company in Indonesia. We migrated their stateless services running on 25+ VMs in the Alibaba cloud to their managed Kubernetes offering (ACK). These application services served Flip's B2B and B2C traffic which often peaked at 25,000 requests/second. Problem Statement. Implement autoscaling of the infrastructure to handle peak traffic. Move to a Cloud-Native architecture to use modern tooling. Solve the problems related to configuration management and operational toil.

Read Case Study

How we saved more than 42% on AWS Infrastructure Costs.
How One2N helped a no-code Shopify platform cut AWS infrastructure costs by 42% without affecting uptime, with savings on compute, RDS, and data transfer.

Read Case Study

Kong API Gateway for Gaming and Sports Betting Integrations
How we used Kong Gateway and Lua to decouple legacy wallet systems and accelerate global expansion for a Tier‑1 gaming enterprise.

Read Case Study

Modernising Observability for a Billion Dollar Revenue Gaming Platform
A gaming and betting operator ran 200+ Java services across 30,000+ VMs on AWS and on-premises, processing 180M spins daily and 600K RPS, but needed unified observability across fragmented monitoring tools.

Read Case Study

Migrating from AWS Glue to Self-Hosted ClickHouse: 70% Cost Reduction for a Fintech Client.
How One2N cut a fintech client's analytics spend from $13K to $4K/month by replacing AWS Glue, DMS, and Athena with a self-hosted ClickHouse stack.

Read Case Study

How we reduced Kafka message latency by 90% and increased throughput 4x for a $300M SaaS platform.
How One2N reduced Kafka message latency by 90% and increased throughput 4x for a $300M SaaS platform, while cutting AWS infrastructure costs by 40%.

Read Case Study

Kong API Gateway for Gaming and Sports Betting Integrations
How we used Kong Gateway and Lua to decouple legacy wallet systems and accelerate global expansion for a Tier‑1 gaming enterprise.

Read Case Study

Modernising Observability for a Billion Dollar Revenue Gaming Platform
A gaming and betting operator ran 200+ Java services across 30,000+ VMs on AWS and on-premises, processing 180M spins daily and 600K RPS, but needed unified observability across fragmented monitoring tools.

Read Case Study

Migrating from AWS Glue to Self-Hosted ClickHouse: 70% Cost Reduction for a Fintech Client.
How One2N cut a fintech client's analytics spend from $13K to $4K/month by replacing AWS Glue, DMS, and Athena with a self-hosted ClickHouse stack.

Blog posts

AI

AWS

Best Practices

CI/CD

Data Engineering

Go

How To

MLOps

Observability

Performance Optimization

Resiliency

Security

SRE

Illustration on a navy background of a terminal running exec 3<>/dev/tcp/host/443, cabled into a network socket, next to a dashed folder labelled /dev/tcp that does not exist

Read Blog

How to test TCP connectivity without curl using bash /dev/tcp

Saurabh Hirani

Principal SRE @One2N

No curl, wget or nc on a locked-down box? Use bash /dev/tcp to check if a port is open and read HTTP endpoints, and learn why it fails in zsh.

Illustration of a wall clock at 2:00 beside versioned policy cards v17 and v18 with a W+1 window token tag, above a stream of events

Read Blog

How temporal versioning prevents configuration drift in stream processing

Kalpesh Deo

Staff software engineer-II @One2N

How temporal versioning with effective-time configuration and window tokens stops configuration drift across tenant time zones, rollovers and event replay.

Read Blog

Why Your EKS Auto Mode NodeClass Keeps Reverting to Default

Ananya Dasgupta

SRE @One2N

How a Secondary CIDR rollout on EKS Auto Mode exposed a configuration ownership conflict between Argo CD and Auto Mode's own controller - and a lesson about which parts of your NodeClass you don't actually own.

Read Blog

OpenSearch Workload Management (WLM): How to Control Heavy Search Queries

Srivatsa RV

SRE @One2N

Learn how OpenSearch Workload Management (WLM) isolates heavy search queries with workload groups, resource limits, and resiliency modes - without over-provisioning.

Read Blog

How to evaluate AI SRE agents for production?

Hemant Kumar

Staff SRE @One2N

AI SRE agents promise faster incident response. The hard part is knowing which ones you can trust when production is noisy. This post shares how we evaluated OpenSRE before we let any of them influence a production incident.

Read Blog

Conway's Law as a lever to fix legacy architectures

Chinmay Naik

CEO @One2N

Conway’s law talks about how software architecture mirrors the org chart. Learn what this means in practice and how to use Reverse Conway Maneuver to refactor existing system architectures.

Read Blog

Hardening Kubernetes Security with Istio Ambient Mode on Flat Networks

Harshwardhan Mehrotra

SRE @One2N

Spandan Ghosh

Content @One2N

Learn how One2N used Istio Ambient mode to enforce zero-trust security on a flat EKS Hybrid network - with mTLS, SPIFFE identity, egress control, and NetworkPolicies.

Read Blog

Git was not built for multiple AI agents. Worktrees fix that.

Srivatsa RV

SRE @One2N

Spandan Ghosh

Content @One2N

Running multiple AI coding agents on the same repo? Git's single-checkout model breaks fast. Learn how git worktrees give each agent its own isolated workspace.

Read Blog

How one parameter stalled Postgres Replication with Debezium for 3 Weeks

Harsh Mishra

SRE

A story about WAL segments, missing events, and the one line of SQL that should have been there from the start.

Read Blog

Multi-Tenant SFTP: How We Built a Zero-Ops Setup on AWS Transfer Family

Kshitish Deshpande

Software Engineer @One2N

Spandan Ghosh

Content @One2N

How One2N built a zero-ops multi-tenant SFTP system for a fintech client using AWS Transfer Family: one endpoint, isolated buckets, no manual server ops.

Read Blog

My new laptop broke Slack. That's how I ended up on Bluefin.

Harsh Mishra

SRE

Spandan Ghosh

Content @One2N

A kernel patch broke Slack on my Dell XPS. Here's how that led me to Bluefin, immutable Linux, bootc, and OCI images and why it solved my problem for good.

Read Blog

Shooting ourselves in the foot by enabling Spring Kafka batching

Mahesh Aravind

SDE @One2N

We enabled Spring Kafka batch listeners to improve throughput and broke non-blocking retries. Here's what went wrong and what actually fixed it.

Read Blog

Why your Architecture should start with Questions, not boxes

Chinmay Naik

CEO @One2N

TL;DR: Architecture should start by asking constraint questions about volume, scale, team skills, budget, and business goals, not by copying what big tech does. Use curious questions ("Help me understand why?") instead of performance questions ("Why didn't anyone think of this?"). Stay with systems long enough to see them fail in production to learn which questions actually matter.

Read Blog

Gitops for Kafka in the real world: How we governed 78 clusters without breaking production

Harshwardhan Mehrotra

SRE @One2N

Spandan Ghosh

Content @One2N

This post is about how we, as a lean platform engg team used gitops and jikkou to bring 78 kafka clusters under control, cut ticket based changes from days to minutes, and kept legacy kafka 0.8.2 clusters safe in production.

Read Blog

Prayogshala - The Engineering Laboratory at One2N

Chinmay Naik

CEO @One2N

We created Prayogshala, One2N’s internal engineering lab, to capture our learnings, experiments and knowledge base. See how it helps engineers get better by asking "why" before "how".

Read Blog

The Gotchas of OTEL collector processors for effective observability in K8s

Sanket Rajgiri

SRE @One2N

Spandan Ghosh

Content @One2N

Struggling to make sense of OpenTelemetry Collector processors for real-world projects? This blog breaks down what each OTEL processor actually does, where it matters, and shares real lessons from messy SRE problems like taming noisy data, surviving crashes, and staying under cost limits in Kubernetes.

Read Blog

How Queueing Theory Makes Systems Reliable

Srivatsa RV

SRE @One2N

Learn how SREs use queueing theory to explain why 70 percent utilisation feels calm, 90 percent feels cursed, and how the same math helps you choose headroom, tame retries, and protect your error budget before incidents hit.

Read Blog

How to read SRE graphs without lying to yourself

One2N

Team

Are your SRE charts messing with your head? We’ll show you step by step how to actually make sense of those dashboards: percentiles, averages, heatmaps, and all, so you spot real issues fast. No jargon, just practical advice from daily SRE work.

Read Blog

Error Budget Calculation: Downtime Minutes for every SLO

Srivatsa RV

SRE @One2N

Turn your SLO into something you can argue about in a meeting: this guide shows how to convert 99.9% into 43 real minutes of downtime, read burn rate, push back on “five nines,” and decide when to ship or hit pause.

Read Blog

Percentiles in SRE: Why averages lie about latency

Srivatsa RV

SRE @One2N

Spandan Ghosh

Content @One2N

See why SREs stop trusting average latency and switch to percentiles, with clear examples of p50, p95, p99, how they expose tail pain for real users, and how to pick the right percentiles for your system.

Read Blog

Comparing Latency vs Throughput: why high utilisation hurts reliability

Srivatsa RV

SRE @One2N

Spandan Ghosh

Content @One2N

This guide explains how latency, throughput, and utilisation really interact, why running at 90 percent capacity quietly wrecks your SLOs, and how to pick safe headroom so your system can survive real-world traffic spikes without falling over.

Read Blog

Deploying a scalable NATS cluster part 2: hands-on demo

Barun Debnath

SRE

Spandan Ghosh

Content @One2N

Learn how to deploy a scalable, production-ready NATS cluster using Docker, complete with observability via Prometheus and Grafana. This hands-on demo covers cluster design, stream replication, JetStream, DLQ, and load testing; helping you build robust messaging infrastructure for cloud native applications at scale.

Read Blog

SRE math every engineer should know: a practical guide

Srivatsa RV

SRE @One2N

Curious how top engineers keep systems reliable? This guide breaks down the math behind Site Reliability Engineering into simple, real-life examples whether it’s understanding error budgets, decoding percentiles, or making sense of dashboards. Perfect if you want to stop firefighting and start making data-driven, confident decisions on call.

Read Blog

From Aircraft to Cloud: my experience completing the SRE Bootcamp

Pankaj Goriya

SRE @One2N

Spandan Ghosh

Content @One2N

From fixing aircraft to building cloud systems, this is my journey through One2N’s hands-on SRE Bootcamp. I share how real-world challenges and team culture helped me grow as an engineer, plus some honest advice for anyone thinking about site reliability engineering.

Read Blog

Debugging KubeVPN's DNS resolution failures

Nilesh D

Software Engineer @One2N

Chinmay Naik

CEO @One2N

The other day, I was trying to access the Go documentation at go.dev, but the page wouldn't load. Not on Chrome, not in incognito mode, not even on Safari. The error was simple but frustrating: "Could not resolve hostname".

Read Blog

What happens when you run Terraform commands?

Sahil Joseph

Staff SRE @One2N

Saurabh Hirani

Principal SRE @One2N

Have you ever wondered what happens when you run Terraform commands? In this post, we will explore the following topics to gain a better understanding of Terraform internals: Architecture and provider plugin architecture, Credential management, State locking and Debugging

Read Blog

Daily driving Omarchy and Hyprland as a CTO

Jaideep Khandelwal

CTO @One2N

A hands-on review of Omarchy Linux, an easy-to-install Arch-based Linux distro with Hyprland window manager pre-configured out of the box. Learn what works, what breaks, and who should use it.

Read Blog

DORA Metrics: Useful or Fluff

Jaideep Khandelwal

CTO @One2N

In this post, we shed light on why following DORA metrics without context might cause more harm than good. We also show how leaders can use a practical, strategic approach to connect DORA metrics to meaningful business outcomes.

Read Blog

A pragmatic guide to get started with OpenTelemetry

Spandan Ghosh

Content @One2N

A guide explaining OTEL, monitoring vs. observability, telemetry’s pillars, OTEL instrumentation, and seamless backend migration. Includes practical migration strategies and trade-offs for SREs adopting OTEL.

Read Blog

How we solved a critical site-to-site VPN IP address conflict in AWS

Rishiraj Rathore

SRE @One2N

Spandan Ghosh

Content @One2N

How One2N solved a critical site-to-site VPN IP address conflict in AWS, including details on the problem, failed solutions, and the final parallel system approach that worked with automation.

Load More

Talks

AWS

Best-Practices

CICD

Cost-Optimisation

Data-Engineering

GO

How-To

IaC

Kafka

k8s

MLops

Monitoring

Networking

Observability

Perf-Optimization

Resiliency

Scalability

Virtualization

Watch Talk

Token Economics, Paisa, Prompts & Production | Entain Tech Brews

Srivatsa RV

SRE @One2N

Biswa Prakash Mishra

SRE @One2N

In this Tech Brews at Entain talk, Srivatsa RV and Biswa dive deep into token economics and the art of token optimization for LLM applications. They explain the fundamental concepts of tokens, the critical differences between prefill and decode phases, and how caching strategies can drastically reduce AI inference costs. By breaking down prompts into cache reads and un-cached tokens, they reveal practical techniques for maintaining cache stability, structuring context, and optimizing dynamic data to achieve up to 90% cache hit rates.

Watch Talk

AI Enabled SDLC | Entain Tech Brews

Jaideep Khandelwal

CTO @One2N

We’ve all “vibe coded” with AI. In this Tech Brews at Entain talk, Jaideep Khandelwal shows how to turn that into a disciplined, AI‑enabled SDLC that teams can trust in production.

Watch Talk

Talking to Coding Agents from Neovim

Saurabh Hirani

Principal SRE @One2N

In this Mitramandal.ai session, Saurabh Hirani goes under the hood of how Neovim talks to CLI‑based coding agents like Claude Code, OpenCode, Pi, and Ko, and why these patterns apply to any editor you use. He traces his own journey from Neovim to VS Code and back, driven by the realization that the best agentic UX often lives on the command line rather than in editor chat sidebars.

Watch Talk

Migrating a Fintech Analytics Stack to Open Source - FOSS Meetup Pune

Kshitish Deshpande

Software Engineer @One2N

How One2N moved a fintech processing 50 to 70M USD daily from AWS Glue and Athena to a self-hosted ClickHouse stack, cutting infra costs dramatically.

Watch Talk

Kubernetes for Hybrid Cloud Environments - #60 Kubernetes Pune Meetup

Harshwardhan Mehrotra

SRE @One2N

How One2N designed an EKS hybrid setup for a betting platform across 40 data centers with strict data residency and latency requirements.

Watch Talk

Platform Engineering Strategy: Scaling from 5 to 500 Engineers | One2N Bits Ep. 6

Jaideep Khandelwal

CTO @One2N

Srivatsa RV

SRE @One2N

When do you actually need a platform team? Srivatsa and Jaideep break down platform engineering from first principles: culture, org size, and adoption.

Watch Talk

Solving Synthetic Data generation using LLMs | mitramadal.ai EP2 - 2026

Chinmay Naik

CEO @One2N

In this talk at mitramandal.ai EP2, Chinmay Naik (Founder & CEO, One2N) walks through how One2N solved the problem of synthetic data generation using LLMs - covering the real-world use cases that drove it, why anonymization falls short, and how they built a tool that uses LLMs for schema generation and deterministic faker-based data generation.

Watch Talk

.zshrc tmuxinator tease | Command Line Heroes EP3 - Bengaluru

Saurabh Hirani

Principal SRE @One2N

Command Line Heroes is a meetup series for developers, SREs, and terminal nerds who love living inside the shell. ​In episode 3 from Bengaluru, Saurabh Hirani (Principal SRE at One2N) walks through a real-world debugging story where Kiro CLI quietly breaks his beloved Tmuxinator setup. ​ This talk is a deep dive into how a “simple” autocomplete feature ended up sitting between Ghostty/VS Code, tmux, and zsh, and why things started to go weird: tmux panes showed commands but refused to execute them, shell startup got slower, and mysterious processes began to appear. ​ Along the way, Saurabh uses LLMs, exec, and process inspection to peel back the layers and explain what’s really happening inside your terminal sessions. CLI Heroes Ep3 was an in person meetup held on 31st Jan 2026, thanks to Smarsh Inc for providing the venue, this is just a small snippet of the things that happened at the meetup. Make sure to follow One2N events on luma to be notified of upcoming events: https://one2n.io/meetups

Watch Talk

The implications of AI in the SRE world | One2N Bits Ep.5

Jaideep Khandelwal

CTO @One2N

Saurabh Hirani

Principal SRE @One2N

AI in SRE is not just about flashy incident bots or code‑completion in your terminal. In this One2N Bits episode, Saurabh Hirani (Principal SRE, One2N) and Jaideep (CTO, One2N) break down what it really takes for AI to be useful in SRE teams, from documentation and RCAs to on‑call culture and leadership incentives. ​ ​They discuss whether AI can actually replace SREs, how to think in terms of “sub‑agents” mapped to SRE workflows, and why organizational context, knowledge sharing, and responsibility ownership matter more than yet another AI tool. ​ What you’ll learn Why AI in SRE is more than a code completion agent, and how to scope AI sub‑agents across incident management, capacity planning, and automation. How to handle knowledge hoarding, fear of replacement, and responsibility diffusion when AI is mandated from the top. Practical ways to use AI to clean up RCAs, keep architecture and docs fresh, and shorten onboarding feedback loops for SREs. How AI can free SRE teams from boilerplate and cognitive load so they can focus on business‑critical reliability work.

Watch Talk

Using LLMs to Improve Incident Response Times | mitramadal.ai - 2025

Jaideep Khandelwal

CTO @One2N

mitramandal is an AI meetup by One2N where engineers share how they've actually integrated AI into their daily development workflows. every developer is solving applied AI differently: some are orchestrating agentic pipelines across MCPs, others are wiring real-time RAG loops into Cursor or Claude, and more are operationalizing multimodal copilots inside CI/CD. we've seen fine-tuning flywheels, retrieval sandboxes, and autonomous evaluation harnesses that ship to prod every week. Talk 1: Using LLMs to Improve Incident Response Times - Jaideep Khandelwal : https://youtu.be/f_nV4GyWO0k?si=Zt_PnLzongzt4E9q Talk 2: The long way to production - Aditya Kasibhatla : https://youtu.be/SWW9rEzgxvA?si=Tr_Z_hz0gAb0JAqd Talk 3: vMCP LEGO blocks for AI workflows and agents - Shiladitya Banerjee : https://youtu.be/Fi8j3NG7WYA?si=xHRJwandjrJLUo0N

Watch Talk

AWS Outage: Real Costs & Disaster Recovery | One2N Bits Ep.4

Jaideep Khandelwal

CTO @One2N

Srivatsa RV

SRE @One2N

Srivatsa and Jaideep break down disaster recovery after a real AWS outage: RTO, RPO, the economics of DR setups, and what cloud promises don't cover.

Watch Talk

Common pitfalls in Data Analytics: Patterns over Tools | One2N Bits Ep.3

Kshitish Deshpande

Software Engineer @One2N

Jaideep Khandelwal

CTO @One2N

Kshitish from One2N breaks down the most common analytics traps startups fall into and why mental models matter more than tool choices.

Load More

Engineering stories

Engineering stories

Read PSES

The database that was quietly eating itself
P50, P90, and P99 alerts keep firing. APIs are slow. You investigate. A slow-burning incident story about a database degrading itself over time.

Read PSES

Debugging a website that worked on every device except mine
Debugging a website that worked on every device except mine. Storytime! Colleague: Hey, can anyone help? I can't access go.dev on my work laptop. Tried different browsers, cleared DNS cache, nothing works.

Read PSES

A curious case of Postgres choosing the wrong index
A core feature goes down for some users during peak traffic. You're on call. A story about Postgres picking the wrong index and what it took to fix it.

Read PSES

Data engineering mystery - rerouting large data in Kafka
You're a tech lead pulled into a Kafka issue by a colleague. A data engineering mystery about rerouting large-scale data without breaking the pipeline.

Read PSES

The Mystery of Failing Database Writes
You're on call when a PagerDuty alert fires for the main transactional database. An SRE story about diagnosing failing writes under pressure.

Read PSES

A story about a nightmare scenario for every SRE
A cloud security failure wakes you up via an AWS billing alarm. A story about why good engineering practices matter during the One-to-N journey.

Read PSES

The database that was quietly eating itself
P50, P90, and P99 alerts keep firing. APIs are slow. You investigate. A slow-burning incident story about a database degrading itself over time.

Read PSES

Debugging a website that worked on every device except mine
Debugging a website that worked on every device except mine. Storytime! Colleague: Hey, can anyone help? I can't access go.dev on my work laptop. Tried different browsers, cleared DNS cache, nothing works.

Read PSES

A curious case of Postgres choosing the wrong index
A core feature goes down for some users during peak traffic. You're on call. A story about Postgres picking the wrong index and what it took to fix it.

Read PSES

Data engineering mystery - rerouting large data in Kafka
You're a tech lead pulled into a Kafka issue by a colleague. A data engineering mystery about rerouting large-scale data without breaking the pipeline.

Read PSES

The database that was quietly eating itself
P50, P90, and P99 alerts keep firing. APIs are slow. You investigate. A slow-burning incident story about a database degrading itself over time.

Read PSES

Debugging a website that worked on every device except mine
Debugging a website that worked on every device except mine. Storytime! Colleague: Hey, can anyone help? I can't access go.dev on my work laptop. Tried different browsers, cleared DNS cache, nothing works.

Read PSES

A curious case of Postgres choosing the wrong index
A core feature goes down for some users during peak traffic. You're on call. A story about Postgres picking the wrong index and what it took to fix it.