Aureus Blog
Infrastructure insights
Engineering deep-dives, product updates, and hosting best practices from the Aureus team.
Running pgvector in Production: One Year of Lessons
A year ago we started seeing pgvector workloads on customer clusters in earnest. Some patterns work beautifully. Others quietly destroy your query planner at exactly the wrong moment. Here is what we have learned.
Four New Addons: PenTest, Staging Clusters, and GPU Infrastructure
We're expanding the addons marketplace with penetration testing, dedicated staging clusters for S and M tiers, and a GPU cluster setup engagement for AI/ML workloads.
From Vibe Coding to Production: The DevOps Gap Nobody Talks About
AI-assisted development has made building production-quality software faster than ever. Getting it to actually run in production — securely, reliably, at scale — is a different problem entirely.
PgBouncer vs Built-in Connection Pooling: What Actually Matters
PostgreSQL 17 ships with improved built-in pooling capabilities. Does that mean you can drop PgBouncer? Usually not. Here's the practical breakdown.
Backups Are Not a Disaster Recovery Strategy
Having backups is necessary. Testing them regularly and knowing exactly how long recovery takes — that's a disaster recovery strategy. Most teams have the former and not the latter.
PostgreSQL 17 in Production: What We Changed and What We Left Alone
We've been running PostgreSQL 17 across our cluster fleet since the stable release. Here's an honest account of what's better, what required tuning, and one gotcha we didn't see coming.
How We Migrated a 2TB PostgreSQL Database with Zero Downtime
Moving terabytes of production data without dropping a single request sounds impossible. Here's the exact playbook we used for one of our largest cluster migrations.