Skip to content
TEGAR-AJA avatar

Blog

Troubleshooting across the full stack

App → OS → storage → network → virtualization → hardware → provider. Real incidents rarely respect layer boundaries.

OperationsNetworkingLinux

When production breaks, the cause is often not where the symptom appears. A slow API might be disk pressure; a 'frontend bug' might be a BGP flap; a 'network issue' might be a full root filesystem.

My incident order: measure impact → preserve data → identify root cause → restore service → validate → document. That sequence works whether it's Nginx, a MikroTik, or a Kubernetes node.

Working across web dev, NOC, sysadmin, and cloud taught me that production systems cannot be understood from only one layer.

Read more

Infrastructure as a product, not a pile of servers

Why Kilat Cloud is built as an abstraction layer — API-driven provisioning instead of manual config.

What managing 40+ Linux servers taught me

Patterns from enterprise/dedicated server operations: monitoring, failover, automation, and out-of-band hardware.

Why I ship static-first when I can

Static HTML is still undefeated for content sites — here's how I decide when to reach for it.

Your GitHub is a portfolio — treat it like one

187 public repos: consistency and good READMEs matter more than any single flashy project.

Accessible by default, not as an audit

Focus states, contrast, and semantics are cheaper to bake in than to retrofit.

Performance budgets that actually stick

Setting a number is easy. Keeping it green takes tooling and culture.

From side project to something people use

The unsexy work: auth, error states, empty states, docs — the gap between demo and product.

← All posts