Kaltheron
Long-form technical notes, written by operators for operators.
The Hidden Cost of Chatty Microservices
July 13, 2026
Decomposing monolithic applications substitutes in-memory execution paths with network boundaries, creating massive fan-out overhead. Where an operation once invoked three internal functions, it now executes three network dispatches across distinct services, multiplying timeout policies and failure surfaces.
Compound probability catches engineering organizations off guard. Even with individual services maintaining three nines of uptime, traversing ten downstream systems drives composite availability beneath ninety-nine percent. External callers perceive the cumulative failure rate, not the isolated health of each microservice.
Reading Latency Percentiles Without Fooling Yourself
August 28, 2026
Arithmetic averages hide the extreme tail latencies that percentiles make evident. Even if an endpoint posts an average duration of fifty milliseconds, one out of twenty calls might experience a grueling two-second delay; customers subjected to cold paths and overloaded database shards are the ones …
Practical Notes on Postgres Connection Pooling
May 15, 2026
Postgres connections are heavy - each backend carries megabytes of memory and a full process - so every scaled-out installation grows a pooler, and every pooler grows a dialect. Session pooling is nearly transparent; transaction pooling is fast and unforgiving, quietly breaking session features like…
What Good Observability Actually Looks Like
May 22, 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.…
When to Choose a Queue Over a Request
September 21, 2026
Direct synchronous communication remains standard practice because execution flow is intuitive: an upstream service queries, the target responds, and delay is immediately measurable. Message queues become necessary when processing duration exceeds HTTP timeouts or when traffic spikes must be leveled…
More reading
- HTTP/3 and QUIC: What Changed for Operators — Networking, May 1, 2026
- Zero-Downtime Deployments Without the Drama — Operations, May 26, 2026
- Why Edge Caching Still Matters in 2026 — Infrastructure, April 24, 2026
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.