Nexarion
Long-form technical notes, written by operators for operators.
Practical Notes on Postgres Connection Pooling
August 18, 2026
Handling PostgreSQL backends imposes substantial overhead because each active client spawns a dedicated process consuming significant RAM, making intermediary connection management mandatory at scale. While session-level pooling preserves compatibility seamlessly, transaction pooling aggressively reassigns connections between queries, breaking features like LISTEN channels, advisory locks, and runtime configuration parameters.
Operational headaches with connection proxies usually stem from poor connection hygiene: forgotten checkouts left dangling in application code, cursors maintained across idle sleep loops, and large migrations hogging pooled slots. Enforcing transaction pooling surfaces these bad patterns immediately by terminating unsupported stateful operations.
A Field Guide to Graceful Degradation
June 28, 2026
Complex platforms require predetermined shedding tiers when capacity degrades. Personalized feeds should drop before transactional checkout; suggestion dropdowns should degrade before primary queries; media thumbnails should omit before raw assets. Documenting and enforcing these boundaries via circ…
The Hidden Cost of Chatty Microservices
August 26, 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 …
Understanding TLS 1.3 Session Resumption
July 21, 2026
TLS 1.3 collapsed the handshake to a single round trip, but the real latency win for repeat visitors comes from resumption. A client presenting a valid pre-shared key can skip straight to encrypted application data, which on mobile networks can be the difference between a page that feels instant and…
Cache Invalidation Patterns That Survive Traffic
August 16, 2026
Engineers constantly repeat the classic computer science joke about cache invalidation, yet few design caches resilient to thundering herds. Request coalescing remains the critical design pattern: upon key expiration, an isolated worker fetches fresh data while concurrent requests receive stale reco…
More reading
- What Good Observability Actually Looks Like — Operations, May 6, 2026
- Why Edge Caching Still Matters in 2026 — Infrastructure, September 12, 2026
- Data Residency Basics for Global Teams — Compliance, September 4, 2026
About us
We cover the unglamorous middle of software: queues that back up, caches that lie, and DNS at 3 a.m. Everything is written from real operational experience.