Traditional NGINX access logs are flat text files. Analyzing them requires awkward awk, grep, and sed pipelines. What if your access logs lived in a proper database instead? The NGINX SQLite logging module (ngx_http_sqlitelog_module) writes access logs directly into SQLite databases. This makes every request instantly queryable with SQL.
Why Use NGINX SQLite Logging?
NGINX’s built-in access_log…
DBeaver 26.0 arriva con una serie di fix importanti per editor, connettività, diagrammi ER e supporto ai database: una release di manutenzione che rende l’esperienza più solida e affidabile. #DBeaver #Software #Linux #Database
Hi! I have one blog on here that’s just been fan-content, but as bio states, this is my scrapbook, and I want to throw something I find both interesting and important in it.
There’s been news about Persona having databases connected to the US Federal Government. If you are interested, read the blog post from one of the researchers here.
The concept of databases is one of the biggest reasons why I am so concerned about ID verification. It doesn’t take much coding experience (if any) to connect a credit card and an email to a message you send on Discord. To explain what I mean, I want to provide an explanation on databases. Please note that this is a very surface level approach to help you understand the basics of databases so I can dig into why I find this so important when it comes to the internet safety debate.
At its most basic, a database is a collection of information stored in tables connected by relationships. You’ve likely seen a table in school or at work. They have rows of data stored based on columns. Here is an example of what a table would look like in a database design:
This table stores a professor’s last name, first name, title, and discipline. There is also a primary key called profID. Primary keys are unique values for each professor. The “data” below would be an example of what would be in the Professors table.
Databases have multiple tables to store more interconnecting data. Let’s say I have a Classes table. Each class needs to have a professor, so I link the Classes table and the Professors table by the professor column in Classes and the profID column in Professors as shown.
The arrow connecting them tells the reader how the tables relate. There are several arrows people use in database structures. Here are the 3 I will be using in this post.
From top to bottom, I’m showing you:
In this case, each class can have one professor, but a professor can have 1+ classes. Let’s see how data in the Classes table will look like.
Note how the professor column has a number instead of a name. That’s because the professor’s ID is connected to the Classes table, not their name.
Speaking of classes, we need some students! I made a Students table with a Many-to-Many relationship with the Classes table between student column in Classes and studentID column in Students
Let’s add some students to the Classes from the previous example.
Just like before, the student IDs are in the student column, not their names.
With my basic lesson on databases complete, here’s why I find age verification so scary. Below is a very oversimplified database for Discord with 6 basic tables that relate to each other.
If they find something you say dangerous, they have access to your contact information/credit card information from Discord and your address/name from your government ID. They can use this information to arrest you.
If allowed by Discord, the government can then look into the data for the server you sent that message in to find other “dangerous individuals” to watch. Once those people submit their IDs, more arrests can be made.
This is violating in so many ways. Without even getting into how your card and purchases can be tracked once they get access to your card number, your privacy and right to speak freely are revoked. You will no longer be able to criticize those in charge. The definition of “dangerous” can be changed at any time to target anyone. It’s dystopian.
I could make an entire post debunking the “for the kids” argument and what kind of policies could be implemented instead if they really want to go that route, but suffice to say, this is not a good solution for anyone except the governments and the big corporations that lobby for policies like this. If you have any questions about this or want to discuss more, let me know! I’d be happy to help the best I can.
Una base de datos NoSQL es un tipo de base de datos que no usa el modelo relacional de tablas y claves.
El nombre NoSQL significa literalmente “Not only SQL” (no solo SQL), lo que indica que pueden existir otras formas de organizar y consultar los datos además del lenguaje SQL tradicional.
Documentales
Clase-Valor
Columnares
Grafos
Una base de datos relacional es un sistema que organiza la información en tablas que se relacionan entre sí mediante claves (key).
La palabra importante es relacional porque los datos no viven aislados:
se conectan mediante relaciones lógicas.
Una base de datos relacional se compone de:
Imagina dos tablas:
Aquí:
Eso permite decir:
El pedido 101 pertenece a Ana.
La relación evita duplicar información innecesaria.
Las bases de datos relacionales:
Antes de este modelo, los datos se organizaban de forma jerárquica o en archivos planos.
El modelo relacional permitió:
En bases de datos NoSQL (como MongoDB), los datos no necesariamente se dividen en tablas relacionadas.
A veces se guardan documentos completos sin normalización.
Una base de datos transaccional es aquella diseñada para registrar y gestionar operaciones que ocurren en el día a día de una organización, garantizando que cada operación se guarde de forma segura, consistente y completa.
Piensa en ella como el “libro contable digital” que no puede permitirse errores.
En este contexto, una transacción es una operación que:
Ejemplos típicos:
Las bases de datos transaccionales siguen las propiedades ACID:
Esto es crítico en sectores como banca, seguros, retail, telecomunicaciones, etc.
[[MORE]]Suelen ser bases de datos relacionales como:
Estas se usan mucho en sistemas ERP, CRM, sistemas bancarios, etc.
OLTP = Online Transaction Processing
(Procesamiento de Transacciones en Línea)
Es el sistema que gestiona operaciones del día a día en tiempo real.
Un cliente paga con tarjeta en una tienda.
Esa operación:
Todo eso ocurre en un sistema OLTP.
OLAP = Online Analytical Processing
(Procesamiento Analítico en Línea)
Es el sistema que analiza datos históricos para tomar decisiones.
El director financiero quiere saber:
Eso no se consulta en el sistema de caja (OLTP), sino en un entorno OLAP (como un Data Warehouse).
Uno ejecuta.
El otro interpreta.
Primero:
Los datos nacen en un sistema OLTP.
Después:
Se extraen, transforman y cargan (ETL) a un entorno OLAP para análisis estratégico.
Base Transaccional (OLTP)
Data Warehouse (OLAP)
Una empresa primero guarda todo en su base transaccional y luego extrae esa información para análisis estratégico.
Imagina una aseguradora:
Luego, para analizar:
Eso ya se hace en un sistema analítico.
—
Un ERP no es una base de datos transaccional.
Un ERP es un sistema de gestión empresarial que utiliza una base de datos transaccional.
Es decir:
ERP = aplicación / sistema
Base de datos transaccional = motor donde se guardan las operaciones
Un ERP (Enterprise Resource Planning) integra procesos como:
Ejemplos conocidos:
Estos sistemas tienen interfaz, lógica de negocio, reglas, permisos, workflows…
Pero por debajo, guardan todo en una base de datos (normalmente relacional y transaccional).
Cuando en un ERP:
Eso genera transacciones OLTP en la base de datos.
La base de datos garantiza:
Algunas empresas usan el ERP directamente para reporting básico.
Pero para análisis más complejos, extraen datos del ERP hacia un entorno OLAP, porque necesitas entender cómo nacen los datos (OLTP) para analizarlos bien después (OLAP).
Local website plentytalent.com was founded in 2014 based on the perceived need for a regional database of emerging talent in all areas of the creative industries. Three years later, expansion plans are in the works through a crowdfunding campaign in association with the World Bank and Caribbean crowdfunding consulting company VisionFunder.
Co-founder Gavin Luke said the website was created using…
Plenty Talent website shows off creative wares (July 23, 2017)
Arabic handwriting is notoriously hard to read. Now, a project called Fihris is using custom-trained AI to decipher the complex Ruq'ah script. We interviewed the team at the Khayrallah Center (USA) about how this tech is unlocking the history of early Arab immigrants to the Americas.
Dancing QT is a combined music database and player application specially designed for dancing schools and equivalent applications. Key features are an easy-to-use interface, fast search capabilities, playlist management, exact pitching and crossfading.
Looking around the open source landscape for a while, I tried to find a music database and player that is suitable for use in a dancing school…
캐시 TTL 60초. 평소엔 아무 문제 없습니다.
60초마다 캐시가 갱신되고, 그 사이 수천 개의 요청은 캐시에서 처리됩니다. DB는 편하게 쉬고 있죠.
그런데 그 60초가 끝나는 바로 그 순간. 문제가 터집니다.
캐시가 만료됐습니다. 첫 번째 요청이 DB에 쿼리를 날립니다.
그런데 그 쿼리가 끝나기 전에 두 번째, 세 번째… 만 번째 요청이 들어옵니다.
전부 캐시 미스. 전부 DB로 직행.
0.3초 동안 5만 개의 동일한 쿼리가 DB를 덮칩니다.
CPU 100%. 커넥션 풀 고갈. 타임아웃. 장애.
캐시는 “있으면 반환, 없으면 DB 조회” 로직입니다.
문제는 “없으면"의 순간이 모든 요청에게 동시에 온다는 겁니다.
첫 번째 요청이 DB에서 데이터를 가져와 캐시에 저장하기까지 보통 50~200ms 걸립니다. 그 짧은 시간 동안 들어오는 모든 요청이 캐시 미스를 경험합니다.
첫 번째 요청만 DB에 가고, 나머지는 기다리게 합니다.
캐시 미스가 발생하면 락을 획득한 요청 하나만 DB를 조회합니다. 다른 요청들은 그 결과를 기다렸다가 공유합니다.
5만 개 쿼리가 1개로 줄어듭니다.
캐시가 만료되기 전에 미리 갱신합니다.
TTL이 60초라면, 50초쯤 됐을 때 백그라운드에서 미리 DB를 조회해 캐시를 갱신합니다. 사용자 요청은 항상 캐시에서 처리됩니다.
만료되는 순간 자체가 없어집니다.
모든 캐시가 동시에 만료되는 게 문제라면, 만료 시간을 흩어놓으면 됩니다.
TTL을 60초 고정이 아니라 55~65초 사이 랜덤으로 설정합니다. 캐시들이 제각각 만료되면서 DB 부하가 분산됩니다.
캐시는 DB를 보호하는 방패입니다. 그런데 그 방패가 내려가는 순간, DB는 무방비 상태가 됩니다.
그 순간을 어떻게 관리하느냐가 시스템 안정성을 결정합니다.
실제로 5만 개 요청이 0.3초 만에 DB를 죽인 사례와 구체적인 코드 구현은 아래 글에서 확인하세요.
One of my coworkers consistently spells “lock” as “loack”. This includes all the derivatives of this word, like loacked, loacking, unloacked, unloacking, etc. The context is mostly database operations, like locking a table or a user.
Having read some Jonathan Strange and Mr Norrell fanfics where the author faithfully replicated 19th century spellings, among them spelling “clothes” as “cloathes”, I’m starting to feel stirrings of a JSAMN fic as an IT department AU.
Beyond the Survey: How Digital Footprints are Redefining Official Statistics
Measuring the Pulse: Real-time Economic Indicators in the AI Era
these two topics came up in mind while searching the title of blog!
Emerging technologies in official statistics is a new trending topic. Data we know is crucial for every type of activity. But how to capture “data” is most important thing. It is like…
ALT
ALT
ALT
ALT“I AM NOT GOING TO LET YOU GET AWAY AGAIN, NOT WHEN I COULD GET PAYBACK!
Worse yet — you’re the sick bastard who took my brother away from me.
SO, I AM NOT GOING ANYWHERE, I’LL BE WAITING FOR YOU — COME OUT AND FACE ME!
You can’t scare me anymore, not when I am so close to finding out what you’ve done to him.”
Modern businesses operate in a fast-moving digital environment where speed, adaptability, and efficiency decide success. Traditional software development often struggles to keep up with rapidly changing requirements, limited developer availability, and rising costs. This gap has accelerated the adoption of low-code platforms, which allow organizations to build and deploy applications faster while reducing technical complexity.
Low-code is no longer just a productivity tool for developers—it has become a strategic growth enabler for businesses of all sizes. From startups to large enterprises, companies are using low-code to innovate faster, optimize operations, and respond quickly to market demands.
A low-code platform provides a visual development environment where applications can be created using drag-and-drop components, configuration, and minimal hand-written code. Instead of building everything from scratch, teams assemble applications using prebuilt modules and workflows.
This approach lowers the barrier to application development, enabling not only developers but also business teams to participate in building solutions that directly address operational needs.
Speed is a competitive advantage. Low-code platforms significantly reduce development cycles, allowing businesses to launch applications in weeks instead of months. This rapid delivery helps organizations respond to customer needs, regulatory changes, and new opportunities without long delays.
By minimizing manual coding and rework, low-code lowers overall development and maintenance costs. Businesses can achieve more with smaller teams and focus skilled engineers on complex, high-value problems rather than repetitive tasks.
Low-code enables closer collaboration between IT and business stakeholders. Domain experts can directly contribute to building solutions, ensuring that applications align closely with real-world workflows and business goals.
Organizations often rely on multiple disconnected systems, spreadsheets, and manual processes. Low-code platforms make it easier to automate workflows and create internal tools that unify data and processes across departments.
This leads to improved productivity, fewer errors, and better visibility into operations—all of which directly support scalable growth.
Applications built on low-code platforms can centralize data and present it through dashboards and reports. When leaders have faster access to accurate information, they can make better strategic decisions that drive business performance.
As low-code adoption grows, security and reliability are no longer optional. Modern platforms are designed with enterprise-grade security in mind, supporting role-based access, authentication, audit logs, and compliance requirements.
In addition, these platforms are built to scale, ensuring applications can handle increasing users and data volumes as the business grows.
Businesses rarely operate on a single system. Low-code platforms offer strong integration capabilities, making it easier to connect applications with databases, cloud services, APIs, and third-party tools.
This connectivity ensures that new applications fit naturally into the existing technology ecosystem rather than creating new silos.
Low-code allows organizations to experiment quickly. Teams can prototype ideas, test them with real users, and iterate based on feedback—all without heavy upfront investment. This culture of experimentation encourages innovation while keeping risks manageable.
These use cases demonstrate how low-code directly supports business agility and growth.
Low-code platforms have evolved into a powerful foundation for modern business growth. By accelerating development, reducing costs, improving collaboration, and supporting secure, scalable applications, they help organizations stay competitive in a digital-first world.
For businesses looking to innovate faster and operate more efficiently, adopting low-code is no longer just an option—it is a strategic advantage.
No. While startups benefit from speed and cost savings, large organizations also use low-code to improve agility, modernize workflows, and reduce pressure on engineering teams.
Low-code complements traditional development rather than replacing it. Complex systems and highly customized applications may still require full-code approaches, while low-code handles rapid delivery and process-driven solutions.
Low-code reduces repetitive development work and enables IT teams to focus on architecture, security, and complex integrations, improving overall efficiency.
Yes. Modern low-code platforms are designed to scale and support growing user bases, data volumes, and evolving business needs.
The Senior Dev Secret: Mastering SQL Execution Plans 💾🔍
Junior developers write queries that work. Senior developers write queries that scale. The difference? Knowing exactly how the Database Engine interprets your code.
The Problem:
You write a simple JOIN. It works fine with 100 rows. But at 1,000,000 rows, your application times out. Most devs just “add an index” and pray. This is not engineering; it’s gambling.
The Solution:
Execution Plans. By using the EXPLAIN command, you pull back the curtain on the Query Optimizer to see exactly where the bottlenecks live.
What to look for in the plan:
🐢 Sequential Scans: The database is reading every single row. This is your #1 performance killer.
🔗 Nested Loops: High-cost joins that multiply your latency exponentially.
📦 Memory Spills: When your query is so inefficient it has to write temporary data to the disk.
Stop guessing. Start reading the roadmap your database gives you. Optimization is about logic, not luck.
👇 ASSETS:
📃 Blog: https://scriptdatainsights.blogspot.com/2026/01/sql-execution-plans-guide.html
🎞 Video: https://youtu.be/aeoKEaDeIy4
🛒 Gumroad: https://scriptdatainsights.gumroad.com/l/january-skills-2026
👇 FOLLOW US:
I have long enjoyed tinkering around the edges of the technology and software world. I haven’t shared much on this site lately other than paintings, but over the past year, I built an app to track my art and some dashboards to share stats. (more in the post) #app #appsheet #dashboard
