Abhay Chaurasiya

Jaeger SPM had static JSON dashboards — hard to maintain, impossible to keep in sync with changing metrics.

Built a programmatic Go SDK dashboard generator that replaces static JSON with code-driven Grafana provisioning.

GoGrafanaSDKObservability

Dynamic metric config across entire monitoring pipeline. Foundation for Jaeger's next-gen observability stack.

CI Sync-Check Workflow

ADR-007 Step 3

#8240

Dashboard generator output could drift from what's committed — no automated way to catch it in CI.

Added a GitHub Actions workflow that validates generator output matches committed dashboards on every PR.

Prevents silent dashboard drift in CI/CD. Enforces consistency across the entire observability pipeline.

GoGitHub ActionsCI/CD

MCP Server Response Limits

jaegermcp reliability fix

#8242

The jaegermcp server had no output size limit — unbounded responses could crash or hang MCP clients.

Enforced response size limits in the MCP server tool handlers, with clean truncation and error surfacing.

Improved reliability of Jaeger's MCP integration. Prevents resource exhaustion in tool response pipelines.

GoMCPReliability

Grafana + SPM Restore

ADR-007 Step 1

#8215

Grafana service was removed from the docker-compose SPM example, breaking the full local observability stack for all users.

Restored Grafana + SPM to the docker-compose example as the first step of the ADR-007 migration.

DockerGrafanaObservability

Unblocked every developer trying to run the full Jaeger monitoring stack locally. First PR in the ADR-007 chain.

HelmCNCF Graduated · 29k+ ⭐

Other Merged PRs

helm/helm · PR #31931

Remove legacy Go import path comments

Cleaned up pre-Go-modules Kythe import path comments from pkg/kube — Go module-based projects no longer require these annotations.

Impact:

Keeps the most widely used Kubernetes package manager's codebase idiomatic and free of legacy noise.

GoCNCFKubernetes

GoReleaser14k+ ⭐

goreleaser/goreleaser · PR #6462

Fix filterOut tag selection bug

Function was returning the excluded tag instead of remaining tags when multiple ignore_tags patterns were configured — breaking release pipelines.

Impact:

Fixed silent release pipeline failures affecting any project using multiple ignore_tags in GoReleaser config.

GoCI/CDRelease

All Merged Pull Requests

4 merged · hover to pause

helm/helm

#31931

pkg/kube: remove legacy import comments

GoKubernetesCNCF

Mar 12, 2026

29k+

goreleaser/goreleaser

#6462

GoGrafanaCNCFObservability

fix: filterOut returns excluded tag when multiple ignore_tags set

feat(monitor): restore Grafana to SPM docker-compose example

feat(monitor): Go SDK dashboard generator — ADR-007 Step 2a

pkg/kube: remove legacy import comments

GoKubernetesCNCF

Mar 12, 2026

29k+

goreleaser/goreleaser

#6462

GoGrafanaCNCFObservability

fix: filterOut returns excluded tag when multiple ignore_tags set

feat(monitor): restore Grafana to SPM docker-compose example

feat(monitor): Go SDK dashboard generator — ADR-007 Step 2a

Mar 24, 2026

20k+

Active Contributions

9 open

gh pr list --author abhay1999 --state open

CACHED

argoproj/argo-cd#26918

fix: add X-Frame-Options and CSP headers to Swagger UI

meshery/meshery#18039

fix: optimize GetSystemDatabase to avoid N+1 queries

NVIDIA/NeMo-Retriever#1645

helm: add nodeSelector and tolerations to NIMCache templates

docker/cli#6861

testing: address G115 integer overflow warnings and re-enable linter

jenkinsci/jenkins#26473

Fix ParserConfigurator to not reference Jenkins class on agents

grafana/grafana#120374

fix(email): remove external remote content from email templates

helm/helm#31933

remove legacy import comments from remaining packages

openvinotoolkit/openvino#34645

GHA(Android): parameterize emulator test artifact names via matrix

Rancheroo/r8s#114

feat: add multi-arch Docker image and GHCR publish workflow

View GitHub Profile

13+ total contributions · growing

$build~/projects›project_manifest.load()

05.

Featured Projects

5 handpicked projects · more on GitHub

self-healing-k8s ~ kubectlDevOps

$ kubectl get pods -n platform

NAME READY STATUS

self-healer-ctrl-x9k2 1/1 Running

prometheus-server-m7p1 1/1 Running

alertmanager-0 1/1 Running

✓ self-healing loop: ACTIVE

Self-Healing K8s Platform

3 custom K8s operators

Problem

Production clusters crash silently at 3am — on-call engineers waste hours manually restarting pods with no root-cause visibility.

Built

Built 3 Go operators using controller-runtime that detect CrashLoopBackOff, trigger automated pod remediation, and restore health without human intervention.

47+Incidents Healed

~11sAvg Heal Time

100%Success Rate

Arch

Custom CRDs · Prometheus webhooks · Helm chart · GitOps

KubernetesGoPrometheusHelmGitOps

dev-env-platform ~ go runDevOps · Go

$ devenv spin up --stack=go-postgres

✓ Namespace dev-go-001: Created

✓ PostgreSQL pod: Running

✓ Go app pod: Running

✓ Port-forward :5432 :8080 ready

● stack ready in 12s

Dev Environment Platform

Single CLI · K8s isolated stacks

Problem

Spinning up a dev stack takes 20+ mins of config, breaks across machines, and can't be reproduced — "works on my machine" is a real blocker.

Built

Go CLI that provisions isolated K8s namespaces with a full stack (MERN / React / Go+PG) in one command — with port-forwarding, teardown, and lifecycle managed automatically.

<15sStack Spin-Up

3Stack Templates

1 cmdFull Setup

Arch

client-go · Dynamic namespaces · Port-forward mux · Go CLI (cobra)

GoKubernetesCLIMERNPostgreSQL

abhay-portfolio ~ nextFull-Stack

$ npm run build

▲ Next.js 14.2.5 (static export)

✓ Compiled successfully

✓ Generating static pages (8/8)

✓ Docker image built

✓ deployed · nginx serving

Developer Portfolio

Live GitHub API · 3D UI

Problem

Generic portfolio sites list skills but don't prove DevOps depth — no interactive demos, no live data, no evidence of system thinking.

Built

Built this site with Next.js 14 + Docker, including a live K8s self-healing visualizer, DevOps playground terminal, CI/CD health monitor, and CNCF PR feed — all pulling live GitHub data.

9Interactive Sections

LiveGitHub API Data

A+Lighthouse Score

Arch

Next.js static export · Docker + Nginx · Vercel · GitHub REST API

Next.jsTypeScriptFramer MotionTailwindDocker

netflix-eks ~ terraformDevSecOps

$ terraform apply -target=module.eks

aws_eks_cluster.netflix: Creating...

aws_eks_cluster.netflix: Complete

✓ Jenkins pipeline: PASSED

✓ Trivy scan: 0 critical found

✓ EKS cluster: RUNNING

Netflix Clone on EKS

0 critical vulns · full CI/CD

Problem

Deploying on AWS EKS manually is error-prone — no security scanning, no reproducible infra, no observability, and configuration drift breaks deployments.

Built

Complete DevSecOps pipeline: Terraform provisions EKS, Jenkins runs CI/CD stages, SonarQube enforces quality gates, Trivy scans images, Prometheus + Grafana provides full observability.

0Critical CVEs

100%IaC Coverage

AutoDeploy on Merge

Arch

Terraform modules · Jenkins pipeline · Multi-stage Docker builds · EKS node groups

AWS EKSTerraformJenkinsDockerTrivy