Articles on Server & Database.

Cloud-native infrastructure, time-series databases, hybrid storage — running 24/7 with committed SLAs and predictable cost.

Server & DatabaseSubmarine cablesRedundancyOperations

Another cable break: check your own routing in 10 minutes, and know when buying more actually helps

On 25 August, three of Vietnam's six submarine cables were down at once. This article is not about redundancy theory — it gives you four commands you can run right now to find out where your traffic goes and whether it is congested, how to tell where your servers sit from a single number, and the real prices of each backup option so you know which one is worth buying.

28/08/202614 min read

Server & DatabaseDisaster RecoveryRansomwareRTO/RPO

A backup is what you have restored, not what you have stored

Most organisations knocked over by ransomware had backups. The failure sits in two places nobody measures: backup repositories are targeted in 96% of attacks, and when the attacker sits quietly inside for months before encrypting, the most recent copy — the thing every backup design optimises for — is the infected one.

08/08/202613 min read

Server & DatabaseMigrationPostgreSQLZero-downtime

Zero-downtime database migration: dual-write, backfill, and a safe cutover for a live system

Swapping a database engine, splitting a schema, or moving cloud providers for a system already serving real customers can no longer be solved with a midnight maintenance window. The four stages of a safe migration — dual-write, backfill, verify, cutover — and the most common trap at each step.

03/07/20268 min read

Server & DatabaseKubernetesMLOpsGPUInfrastructure

Kubernetes for ML and data workloads: designing a cluster from node pools to GPU scheduling

Kubernetes is the standard for container orchestration — but default configurations don't hold up under ML workloads. GPU nodes need dedicated taints and tolerations; training jobs can starve the entire cluster without resource quotas; model serving needs PodDisruptionBudgets to avoid mid-inference interruptions. Four common problems when running ML on Kubernetes and how to design the cluster correctly from the start.

19/06/20269 min read

Server & DatabaseTimescaleDBClickHouseIoT

Database architecture for IoT systems: when to use TimescaleDB, when to use ClickHouse

The two most popular time-series engines for sensor data — selection criteria, schema design, and a data lifecycle strategy so costs don't explode with device count.

08/06/20267 min read

← See all articles