Matej Jelluš Blog
One of the best ways to learn something new is to actually try it. I wanted to learn how to use AWS and some of its services, so I moved my projects to AWS. This was just a very basic setup (EC2, S3, CloudFront). Later I wanted to use additional services like Cognito, RDS, MediaConvert, Simple Queue Service, Simple Email Service, etc. So I started building a new project where I was able to use…
In one of my previous blog posts, I wrote about how to Extend logs in Node.js with unique trace ID. It is a great starting point to improve logging and tracing of your application. But... can it be upgraded to somehow connect the trace IDs into a usable diagram?
User management is part of nearly every application and it is not easy to do it in a proper secure way. One of the possibilities is to use an existing solution such as AWS Cognito.
It is easy to say that someone is not good or weak. But it is not easy to make the decision to fire him / her and it is even harder to tell him / her.
When number of requests to an application grows, it might be hard to read logs and group them together to one single request. This can be solved with trace ID that is unique for each request.
Learning new technologies goes always better if you have something to test it on. When I wanted to start using AWS I decided to move there my blog (you are currently reading). And these are the steps it takes.
Container should be able to perform only a very limited set of operations and it is highly recommended to use different user from root. To achieve this both docker file and kubernetes config must be changed.
Write simple Node.js backend using TypeScript, Express and Terminus and run it in Digital Ocean Kubernetes cluster. In the end you will have application which is running in multiple replications and has access to environment variable passed to the application by Kubernetes secret.
This week, 1 - 4 November, Web Summit conference held in Lisbon, Portugal. Me and two other colleagues from Kontentino were there to present Kontentino and listen to some speakers.
A brief tutorial how to create a Node, TypeScript and GraphQL server from scratch with MySQL database. The ouput is a very basic GraphQL server, which you can run, send a simple query to it and get a response. The whole project is available on GitHub.
On couple of my projects, I started using GitHub Actions. I also wanted to use it for pentest-tool-lite, where I want to run TypeScript lint after each push and publish it to npm after release is created. I had some problems which I recently solved, so I am sharing my solution.
Currently I have one full time job and in my free time I am working on two StartUps. One of them is based in US and the second one is in Slovakia. Both are the same size (3 - 5 people), main project is a web application, but the approach is different.
Couple of weeks ago I came across to this interesting rule and it is worth to think about it more deeply.
When you want to create codecoverage report with phpunit, you need to have XDebug installed otherwise you get just error "No code coverage driver is available". I successfully installed it on Windows but I had lot of issues installing it on MacOS.
Git is something that most of us use daily, though there are functions / commands you do not use but still may help in some situations.
Collection of JavaScript examples when it behaves...weird? Funny?
Have you ever heard about yarn resolutions? It is not something you would use everyday, but it is definitely useful.
The first thing you read about setState - it is asynchronous. I understand it but still did not realize a couple of things.
It was a great idea to use git and create portal like GitHub and share source code among other developers. One of the benefits is, you do not need to reinvent the wheel and just use package you need. And because lot of users will add it to their projects, it gets well tested. But I can see there a big problem: maintenance.
You commited something and right after that you saw you forgot to remove log or add another file?