Lunarex

Lunarex

Field notes on building and running reliable internet services.

Infrastructure

Why Edge Caching Still Matters in 2026

July 13, 2026

Periodic declarations that edge caching is dead consistently overlook networking realities: fiber lines still obey physical constraints. Even with cheap bandwidth and fast runtimes, a round trip between Sao Paulo and Frankfurt easily consumes hundreds of milliseconds. No amount of runtime tuning can outrun the wire.

Today's shift centers on the nature of cached payloads rather than the need for them. Modern architectures place dynamic API payloads and HTML shells at the edge using stale-while-revalidate patterns and tight expiration windows. The operational hurdle remains invalidation: purge mechanisms must propagate near-instantly, supported by surrogate keys or cache tags to handle complex dependency graphs.

Continue reading →

Operations

What Good Observability Actually Looks Like

August 11, 2026

Unchecked dashboard proliferation creates visual noise without resolving production emergencies. Effective operational insight functions backward from triage: when an alert triggers and waking staff investigate, instrumentation must pinpoint modifications immediately.…

Security

Managing Secrets Without Losing Sleep

July 5, 2026

There are exactly two ages of secrets management: 'we keep them in an encrypted file' and 'we were audited'. The distance between them is covered by rotation policies, access trails, and the gradual realisation that humans should read production credentials roughly never.…

Engineering

The Operator's Guide to Load Testing

April 10, 2026

Load tests fail to predict production for a consistent reason: the traffic shape is wrong. Uniform random requests against one endpoint tell you about that endpoint. The production killer is the burst of clients retrying in sync after a thirty-second blip, or the cache-cold crawl at 5 a.m. when the …

Operations

Multi-Region Failover Planning

May 25, 2026

Multi-region failover is mostly decided before the incident. The two questions that matter - how fresh does the standby data need to be, and who is allowed to press the button - sound managerial, but they drive every technical choice downstream from replication topology to health-check placement.…

More reading

About us

We are a small team of infrastructure engineers and technical writers. We publish what we learn running production systems: incident retrospectives, protocol deep-dives and tooling notes.

More about the project →