Beyond Kafka and Redis, Part 2: An AI Chat Backend on NATS 2.15

·7 views

In part 1 I walked through the NATS building blocks one at a time: subjects, request-reply, JetStream streams, the Key-Value store, and the Object Store. That post answered "what can NATS replace?" This one builds a real backend with it. The system is the backend for an AI chat product. It has a GPU work queue, token streaming to the browser, conversation history, per-tenant usage metering, run timeouts, and worker ownership of conversations. In a typical stack that's Kafka or SQS for the queue, Redis for counters and locks, a cron service for timeouts, Postgres for history, and a WebSocket ga

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

GeminiTTS: Gemini 3.8 TTS Online for Single-Voice and Two-Speaker Dialogue

If you have ever needed a quick voiceover for a demo video, a podcast intro, or a language-learning clip, you know the usual options: record it yourself, hire a voice actor, or wrestle with a robotic-sounding TTS engine. I wanted something faster and more expressive, so I have been using GeminiTTS, an online text-to-speech studio powered by Gemini 3.8 TTS. GeminiTTS is a browser-based tool: paste your script, choose a voice, and generate audio you can play back or download as WAV. There is nothing to install and no API key to manage. Key features: Two models – Gemini 3.8 Flash for expressive,

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Choose Cron Healthchecks over App Metrics — Safer Missed Cohort Rollbacks

A tenant-cohort experiment is only reversible while the team can tell, quickly and unambiguously, that a scheduled task stopped completing. Short answer: use a deadline heartbeat as the rollback trigger for missed runs, and keep application metrics as the diagnostic layer. Choose metrics alone only when the scheduler itself can reliably emit an absence signal and the alert query preserves tenant, cohort, region, and experiment-revision boundaries. That split matters for an e-commerce SaaS rollout. A new recommendation model might look healthy at the HTTP layer while its nightly catalog-enrichm

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Kyverno Policy as Code no Kubernetes

Kyverno (do grego "governar") é um motor de políticas nativo do Kubernetes, projeto da CNCF, que permite validar, alterar e gerar recursos do cluster usando apenas YAML, sem aprender uma linguagem nova. Em clusters compartilhados, qualquer pessoa com acesso ao kubectl apply pode criar um Pod privilegiado, uma imagem sem tag ou um Service exposto sem querer. Revisão manual não escala, e documentação de boas práticas não impede ninguém de errar. O Kyverno resolve isso aplicando regras no momento em que o recurso chega à API do Kubernetes, antes de ele ser persistido no etcd. O diferencial frente

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Push notifications on iOS without Firebase: talking to APNs directly from Laravel

Most push tutorials for iOS start with "add the Firebase SDK to your app". It works, but it means shipping Google's SDK inside your iOS binary just to forward a message to Apple, and keeping a second service configured for every app. Your Laravel backend can talk to Apple Push Notification service (APNs) directly. All it takes is a .p8 key, a short-lived JWT and one HTTP/2 request. This post walks through the whole thing, including the two mistakes that cost the most time: the wrong token type and the wrong environment.

0 comments Read more at DEV Community