Production Bugs

How to build a hallucination-free chatbot over a complex SaaS database (NL → SQL) that strictly follows business rules?

by

Summary Building a hallucination-free Natural Language to SQL (NL2SQL) chatbot in a complex SaaS environment requires treating the Large Language Model (LLM) strictly as a routing and translation engine, not a decision-maker. The production-grade approach involves an intent-centric architecture where the backend exposes strictly defined “capabilities” or “query templates,” and the LLM’s role is limited … Read more

How to integrate PraxiLabs Virtual Lab simulations into a custom LMS via LTI or API?

by

Summary This postmortem analyzes the integration of third-party interactive virtual lab simulations (specifically PraxiLabs) into a custom Learning Management System (LMS). The core challenge involves establishing a seamless, data-rich connection that feels like a native feature to the user. The integration requires bridging the gap between the LMS and the simulation provider using standard protocols … Read more

Struggling to conceptualize Look At Matrix

by

Summary In 3D graphics, we typically move the world to the camera, not the camera to the world. The glm::lookAt function constructs a view matrix that transforms vertices from world space to camera space. To achieve this, the matrix must perform the inverse translation and rotation of the camera’s actual orientation. The confusion arises because … Read more

jitsi meet self hosted video is disabled to save bandwidth

by

Summary The issue described is not a bandwidth or CPU resource constraint, but rather a network transport failure causing UDP packet loss. The Jitsi Meet system interprets 4–18% packet loss as a degraded connection and proactively disables video streams to preserve the audio channel and session stability. The root cause lies in the VPS network … Read more

Couldn’t deploy Next.js app with pnpm workspaces to Firebase app hosting

by

Summary This postmortem details a deployment failure of a Next.js application within a pnpm workspace to Firebase App Hosting. The build process encountered two distinct phases of failure, leading to an incomplete build artifact and deployment termination. The root cause was a combination of a misconfigured build command order and missing Prisma client generation before … Read more

Should resume skills be stored as JSON, BLOB, or normalized tables for analytics and search performance?

by

Summary Technical Postmortem: Resume Skills Storage Architecture for Analytics and Search The central issue is a database schema anti-pattern where skills are stored as a BLOB (binary large object) in a MySQL user_data table. This approach fundamentally breaks analytics, search performance, and scalability. The root cause is treating semi-structured data (skills) as an opaque binary … Read more

Product price parsing on Wildberries using Kotlin (In the Android app)

by

Summary The core issue is an attempt to perform client-side web scraping of a modern e-commerce platform (Wildberries) directly from an Android application. The user encountered immediate request blocking, likely due to sophisticated bot detection, WAF (Web Application Firewall) rules, and TLS fingerprinting. The root cause is a fundamental architectural mismatch: treating a complex, security-hardened … Read more

How can I get historical pod start/end times and CPU/memory requests from Prometheus (OpenShift/kube-state-metrics)?

by

Summary This postmortem addresses a common operational gap: reconstructing historical pod lifecycles (start/end times) and resource requests in Kubernetes/OpenShift clusters using Prometheus and kube-state-metrics (KSM). The core issue is that while KSM provides real-time snapshots, historical queries for ephemeral resources require specific PromQL patterns and an understanding of metric cardinality lifecycles. The primary failure mode … Read more

Adding an inline button with the ability to open a chat with a user knowing their username

by

Summary A developer attempted to create an inline button in a Telegram bot that would open a direct chat with a user using only their username. The implementation used a URL button pointing to https://t.me/Username. However, the button click resulted in an error message stating the user does not exist, despite the username being valid. … Read more

Link statically.io Update from /main/ to @main

by

Summary Key takeaway: The statically.io CDN endpoint path /main/ was replaced with @main/ by the provider, breaking all existing references in code and websites. This is a classic dependency path change that requires an immediate update to the URL string in the source code. Action required: Search and replace all occurrences of cdn.statically.io/…/main/ with cdn.statically.io/…/@main/. … Read more