Posts

How we reduced cloud monitoring costs by over 50% on GKE by architecting a lightweight, highly available Prometheus + Thanos cluster instead of managed Prometheus.

How we built near real-time CDC data pipelines using MongoDB change streams to keep Elasticsearch in sync across high-volume production collections without performance penalties.

How we evaluated Longhorn for RWX shared storage in Kubernetes, diagnosed stability bottlenecks in dynamic clusters, and migrated to a hardened off-cluster NFS solution.

An introduction to WSGI and Gunicorn internals to understand how Python web servers handle requests and how to debug overlooked production issues.

A practical guide to worker types in Gunicorn: understanding concurrency models, synchronous vs asynchronous workers, and selecting the right worker architecture.

How we implemented request-body-aware dynamic routing in Ingress Nginx using custom Lua plugins to simplify complex microservice migrations without downtime.

How we designed and automated on-demand ephemeral staging environments in a single Kubernetes cluster to enable reliable pre-release integration testing across microservices.

Instead of spending hours on setting up a dev env in your local machine just for a single debug, do it inside a container and delete the container and image after the job. Zero Bloat.