Wittyden

Wittyden

Long-form technical notes, written by operators for operators.

Networking

Structuring DNS for Reliability

July 19, 2026

Nameserver outages are devastating because they paralyze systems globally: infrastructure monitors report operational green across every host while end users cannot locate your endpoints. Mitigating this catastrophic blind spot requires standard multi-provider redundancy backed by distinct network backbones.

Designing cache lifetimes requires careful trade-off analysis. Overly aggressive short TTLs increase resolver query volume and amplify upstream blips, whereas prolonged caching windows shelter clients during outages but immobilize scheduled traffic migrations. Differentiating durations by record volatility—short timers for active endpoints, extended timers for static assets—provides balance.

Continue reading →

Data

Cache Invalidation Patterns That Survive Traffic

May 7, 2026

Everyone quotes the two hard things joke; fewer people ship caches engineered against stampede. The pattern that matters most is request collapsing: when a hot key expires, exactly one worker rebuilds while everyone else serves slightly stale data, which turns a cache miss from a database incident i…

Infrastructure

Why Edge Caching Still Matters in 2026

August 17, 2026

Every few years someone declares the edge cache obsolete: bandwidth is cheap, compute is fast, so why bother? Yet p99 latency keeps telling a different story. The round trip from a user in Sao Paulo to an origin in Frankfurt costs around 200 ms on a good day, and no amount of application optimisatio…

Engineering

When to Choose a Queue Over a Request

May 1, 2026

Traditional RPC calls remain popular due to straightforward causality: the client makes an invocation, waits for the response, and monitors latency directly. Asynchronous message queuing becomes essential when background tasks outlast active connections or when sudden volume surges threaten to overw…

Engineering

The Operator's Guide to Load Testing

August 19, 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 …

More reading

About us

Our contributors have spent years on-call for large platforms. This site collects the playbooks, postmortems and reference material we wish someone had handed us earlier.

More about the project →