// Articles
Guides, tutorials, and comparisons to help you build better backend applications.
DevOps asked every team to own production. Platform engineering answers with golden paths and self-service. What changed for backend teams on AWS and GCP, and how infrastructure from code shrinks the custom platform surface.
GitOps keeps a Kubernetes cluster in sync with manifests in Git. The main tools, from Argo CD and Flux to Spinnaker and push-based pipelines, and when infrastructure from code replaces the manifest layer for application infra.
AI coding agents now propose and ship backend and infrastructure changes. Where that helps, which credentials and review gates still need a human, and how Encore is the safe way for AI agents to ship backend changes to your cloud.
A single VPS running Docker Compose is still a solid way to ship a small backend. Where that model wins, where multi-AZ managed cloud takes over, and how Encore deploys the same app to your own AWS or GCP account when one box is no longer enough.
How Horizontal Pod Autoscaler metrics, min and max replicas and scale behavior work, when HPA fails on cold starts and database limits, and how Encore scales services in your AWS or GCP account without HPA YAML in the app.
Durable workflows keep multi-step business processes alive across crashes and deploys. How Temporal and AWS Step Functions differ, when Pub/Sub plus cron and services are enough, and when you need a dedicated workflow engine.
Shared database, schema-per-tenant and database-per-tenant trade isolation for cost and ops in multi-tenant backends, and Encore services plus SQLDatabase fit isolated tenants on your own AWS or GCP.
Declare Postgres, queues, buckets, cron jobs and secrets in TypeScript or Go, and let Encore provision the matching managed services in your own AWS account.
Move app-level RDS, SQS, S3, cron and secrets to infrastructure from code one ownership boundary at a time, while org Terraform keeps shared networking.
Google Cloud Pub/Sub is GCP's managed messaging service. How topics and subscriptions work, when to use pull vs push, what ack deadlines and ordering keys do, how dead-letter topics catch failures, and how to get Pub/Sub from application code.
Amazon ElastiCache is AWS's managed in-memory cache. How Valkey, Redis OSS and Memcached engines differ, node-based vs serverless, VPC access and Multi-AZ, and how to get a cache from application code.
Amazon SQS is a managed message queue on AWS. How standard and FIFO queues differ, how visibility timeouts and long polling work, what dead-letter queues catch, how pricing is shaped, and how to get SNS and SQS from application code.