Archive

Dive into a collection of past posts and rediscover timeless content.

DevOps

How Great Engineers Debug Production Problems

Oct 5, 2026

•

7 min read

How Great Engineers Debug Production Problems

Boring topic. Extremely useful skill. Read this.

Yoshik Karnawat

Digest

#38 Uptime Sync: 100TB Saved, PostgreSQL Migration Traps, and AI Agent Security

Oct 2, 2026

•

1 min read

#38 Uptime Sync: 100TB Saved, PostgreSQL Migration Traps, and AI Agent Security

From Cloudflare’s Rust memory optimisation to failed Postgres migrations, Kubernetes rollout debugging, and the DNS weakness behind an OpenAI jailbreak

Yoshik Karnawat

Linux

Sep 29, 2026

•

6 min read

Linux Memory: Page Cache, Buffers, and Available RAM

A practical guide to free -m, page cache, reclaimable memory, swap, pressure, and container memory limits.

Yoshik Karnawat

Digest

#37 Uptime Sync: AWS Outages, Postgres Replication, and Cutting Observability Costs

Sep 26, 2026

•

1 min read

#37 Uptime Sync: AWS Outages, Postgres Replication, and Cutting Observability Costs

What a year of AWS failures reveals about multi-region architecture, why DuckDB 2.0 is faster, and how teams are reducing monitoring bills without losing visibility

Yoshik Karnawat

Cloud

Sep 22, 2026

•

6 min read

Choosing a Cloud Region: Latency, Cost, and Resilience

The closest region is not automatically the right region. The real question is whether your workload actually fits there.

Yoshik Karnawat

Digest

#36 Uptime Sync: Growing Past Senior, Cutting Cloud Waste, and Preventing Production Incidents

Sep 20, 2026

•

1 min read

#36 Uptime Sync: Growing Past Senior, Cutting Cloud Waste, and Preventing Production Incidents

Career progression, architecture trade-offs, 100 TB of saved RAM, and practical DevOps lessons for building safer and more efficient systems

Yoshik Karnawat

Networking

How Internet Routing Recovers From Network Failures

Sep 14, 2026

•

6 min read

How Internet Routing Recovers From Network Failures

BGP can move traffic to another path, but the result depends on failure detection, local policy, and whether the backup route is actually usable.

Yoshik Karnawat

Digest

#35 Uptime Sync: Cloud Availability Risks, PostgreSQL 19, and AI Agent Safety

Sep 11, 2026

•

1 min read

#35 Uptime Sync: Cloud Availability Risks, PostgreSQL 19, and AI Agent Safety

The hidden failure modes in cloud software, the infrastructure changes worth tracking now, and why human judgment still matters in AI-powered engineering

Yoshik Karnawat

DevOps

DNS Changed. Some Users Still Reaching the Old Service?

Sep 8, 2026

•

5 min read

DNS Changed. Some Users Still Reaching the Old Service?

The DNS record may be correct, but cached answers, internal DNS policies, or routing can still send users to the old destination.

Yoshik Karnawat

Digest

#34 Uptime Sync: AI Outages, Buildkite Lessons, and the Cloud Costs Hiding in Plain Sight

Sep 5, 2026

•

1 min read

#34 Uptime Sync: AI Outages, Buildkite Lessons, and the Cloud Costs Hiding in Plain Sight

From chatbot downtime and crawler traffic to inconsistent VM performance, safer recovery workflows, and practical GitOps tools

Yoshik Karnawat

Archive

Dive into a collection of past posts and rediscover timeless content.

How Great Engineers Debug Production Problems

#38 Uptime Sync: 100TB Saved, PostgreSQL Migration Traps, and AI Agent Security

Linux Memory: Page Cache, Buffers, and Available RAM

#37 Uptime Sync: AWS Outages, Postgres Replication, and Cutting Observability Costs

Choosing a Cloud Region: Latency, Cost, and Resilience

#36 Uptime Sync: Growing Past Senior, Cutting Cloud Waste, and Preventing Production Incidents

How Internet Routing Recovers From Network Failures

#35 Uptime Sync: Cloud Availability Risks, PostgreSQL 19, and AI Agent Safety

DNS Changed. Some Users Still Reaching the Old Service?

#34 Uptime Sync: AI Outages, Buildkite Lessons, and the Cloud Costs Hiding in Plain Sight

The Uptime Engineer