Recent
Cloud-native monitoring with Lightweight Prometheus + Thanos on GKE
·1022 words·5 mins
How we reduced cloud monitoring costs by over 50% on GKE by architecting a lightweight, highly available Prometheus + Thanos cluster instead of managed Prometheus.
Mongodb triggers and change streams
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.
Rwx Volume in k8s: Adventure with Longhorn
·1001 words·5 mins
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.
Guide to worker types in gunicorn
·953 words·5 mins
A practical guide to worker types in Gunicorn: understanding concurrency models, synchronous vs asynchronous workers, and selecting the right worker architecture.
Introduction to WSGI and gunicorn
·1020 words·5 mins
An introduction to WSGI and Gunicorn internals to understand how Python web servers handle requests and how to debug overlooked production issues.