Engineering decisions.
Explained in practice.
Practical guides to reliable systems, from tenant isolation and integrations to AI retrieval and product operations.
Page 6 of 7 · Articles 61–72 of 84

Salla App Functions: choose the right execution boundary
A practical architecture for separating blocking decisions from background commerce workflows in Salla App Functions.

Durable AI agents: checkpoints are not enough
A production architecture for AI agents that survive restarts, pause for approval and execute external side effects safely under retries.

Salla's light order contract: migrate without losing fulfillment data
A practical migration architecture for Salla order integrations after expanded responses ended, using selective hydration, queues and sequential reconciliation.

PostgreSQL 19 Beta 4: test the upgrade, not the feature list
A practical pre-production test plan for PostgreSQL 19 Beta 4, covering reverted features, REPACK concurrency, replication, query plans and rollback evidence.

Zid multi-location inventory: choose the right batch boundary
A practical architecture for syncing ERP or WMS stock with Zid, choosing product-centric or location-centric batches and verifying every write through reconciliation.

Google Cloud X5 reaches 48 TB: when scale-up is worth the trade-offs
Google Cloud's new X5 bare-metal series raises single-node memory to 48 TB. Here is what architects must test across NUMA, Hyperdisk, recovery and SAP HANA before scaling up.

Salla and ERP inventory sync: make every stock movement auditable
A practical architecture for syncing ERP stock with Salla using bulk deltas, branch mapping, change reasons, a durable outbox and scheduled reconciliation.

GKE raises the limit to 512 Pods per node—should you use it?
GKE Standard now supports 512 Pods per node. Here is the IP math, density benefit, failure impact and a safe node-pool migration plan.

Salla conditional webhooks: filter events before they reach your app
A practical guide to Salla Rules: choosing webhook conditions, registering them safely, testing changes and preserving reliable downstream processing.

AlloyDB for AI agents: live data without sharing production compute
Google's new AlloyDB preview gives AI agents read-only PostgreSQL nodes on isolated compute while keeping an up-to-the-second view of production data.

Salla and Zid webhooks: order sync that survives missed events
A practical architecture for Salla and Zid apps that combines verified webhooks, a durable inbox, idempotent processing and API reconciliation.

Cloudflare Containers: the storage boundary behind tenant isolation
A fresh Cloudflare disclosure prompts a practical review of storage reuse, ephemeral disks and tenant boundaries in SaaS and AI agent infrastructure.
