Berlin, Germany

Puneet Bharti

_

Building resilient cloud-native infrastructure and operating blockchain validators on Kubernetes — Cosmos, Canton Network, Ethereum, and Mina. If it runs in production, it belongs in code.

View Work Read WritingContact Me
0/minRequests handled
0/dayLogs ingested
0Manual work eliminated
0Logging MTBF

About

I've been building and operating infrastructure that doesn't keep me up at night — from bare-metal servers in Noida to multi-region cloud platforms in Berlin.

Today I work at Finoa GmbH, where I design and operate Cosmos blockchain validator infrastructure on Kubernetes. My work philosophy is simple: if you're doing it twice, automate it; if you can't observe it, you don't own it.

India → UAE → Berlin. Practitioner. IaC-first. Obsessed with reducing MTTD and MTTR through better observability.

Full story

Projects

Selected Work

Cosmos Validator Onboarding Automation

Validator onboarding time reduced from days to hours; zero single points of failure in key management.

KubernetesHelmHorcruxCosmos SDK

Unified Logging System (ELK + Kafka)

600 GB/day ingested reliably, ~300 hr MTBF on the logging cluster, sub-second search across all services.

ElasticsearchLogstashKibanaKafka

Multi-Region Hybrid Video Delivery

5% reduction in end-to-end RTMP latency for Asian audience; 99.9% stream availability across regions.

AWSAlicloudNginxTerraform
All projects

Stack

Tools I Work With

Cloud

AWSGCP

Containers & Orchestration

KubernetesDockerHelmcdk8s

Infrastructure as Code

TerraformAnsible

Observability

PrometheusGrafanaELK StackKafka

CI/CD & Networking

GitLab CIJenkinsNginx

Blockchain

Cosmos ValidatorsHorcrux

Languages & Scripting

PythonNode.jsBash

Writing

Latest Posts

15 September 20242 min read

How We Automated Cosmos Validator Onboarding with Helm

Onboarding a new Cosmos validator used to take days of manual YAML and key ceremony. Here's how we reduced it to a single Helm install, with Horcrux threshold signing built in from the start.

KubernetesHelmBlockchainCosmos
10 June 20243 min read

Running ELK at 600GB/day — What We Learned

We ran a production ELK stack ingesting 600 GB of logs per day with ~300 hour MTBF. Here's what worked, what didn't, and what we'd do differently.

ELKKafkaObservabilityAWS
22 March 20243 min read

cdk8s vs Terraform — A Real-World Comparison

I've used both tools in production. They solve overlapping problems differently. Here's when I reach for each, and where the lines blur.

KubernetesTerraformcdk8sIaC