I have tried going through log4js but do we need to create an api to store logs in log4js.
If you're comfortable with Javascript, the simplest solution would probably be to write a little Node web server (using Express, for instance) that you could send errors to when they happen? The server could append them to a file as it receives them.
I've been working on the front-end so far, now I'm going to create my first full-stack application. I want to use node.js, express and AWS for this.
At the design stage, I already encountered a few problems. Therefore, I have a few questions and I am asking you for help:
Can I send a message (simple JSON or database value) from the server to all clients who have already opened my home page in a simple and cheap way?
I'm not talking about logged in users, but all who downloaded the main page (GET, '/')?
Using the admin panel ('www.xxxxxxxxx/admin'), I want to send a message to the server once a day. Then I want to change the HTML to display this message. I was thinking to use EJS for this and download this message from the database.
Can I make it better? If someone visits my home page (GET, '/'), EJS will download the message from the database each time! Even though its value is the same for 24 hours. Can I get the value once and then use it until the value is changed? How to store the message? As a JSON on the server? Or maybe in the .env file?
If the user refreshes the page, do I have to pay for calling all AWS functions to build the page each time? Even if nothing has changed in the files?
How to check if the page has new content and then send it to the user, instead of sending the unchanged page files: .html, .js, .css, etc.?
Can I send the user only the changed, dynamically created html file, and not send again unchanged .js and .css files?
Does every user who opens the home page (GET, '/') create a new connection to the server using WebSocket / socket.io?
I will try to answer some of your questions:
Can I send a message (simple JSON or database value) from the server to all clients who have already opened my home page in a simple
and cheap way? I'm not talking about logged in users, but all who
downloaded the main page (GET, '/')?
I guess you mean sending push notifications from the server to the user. This can be done with different services depending on what are you trying to build.
If you are planning to use GraphQL, you already have GraphQL subscriptions out of the box. If you are using AWS, go for Appsync, which is the AWS service for GraphQL.
If you are using REST and a WebApp (not a mobile app), go for AWS IoT using lambdas. Here is a good resource using Serverless Framework (API Gateway + lambdas + IoT) for unauthenticated users: https://www.serverless.com/blog/serverless-notifications-on-aws
If you are planning to use notifications on a mobile app, you can go for SNS, the "de facto" service for push notifications in AWS world.
Using the admin panel ('www.xxxxxxxxx/admin'), I want to send a message to the server once a day. Then I want to change the HTML to display this message. I was thinking to use EJS for this and download this message from the database. Can I make it better? If someone visits my home page (GET, '/'), EJS will download the message from the database each time! Even though its value is the same for 24 hours. Can I get the value once and then use it until the value is changed? How to store the message? As a JSON on the server? Or maybe in the .env file?
Yes, this is the way it's expected to work. The HTML is changed dynamically using frontend code in Javascript; which makes calls (using axios for example) to the backend every time you get into, i.e. "/" path. You can store this data in frontend variables, or even use state management in the frontend using REDUX, VUEX, etc. Remember the frontend code will always run in the browser of your users, not on your servers!
If the user refreshes the page, do I have to pay for calling all AWS functions to build the page each time? Even if nothing has changed in the files?
What you can do is store all your HTML, CSS, Javascript in an S3 bucket and serve from there (this is super cheap, even free till a certain limit). If you want to use Server Side Rendering (SSR), then yes, you'll need to serve your users every time they make a GET request for example. If you use lambda, the first million request per month are free. If you have an EC2 instance to serve your content, then a t2.micro is also free. If you need more than that, you'll need to pay.
How to check if the page has new content and then send it to the user, instead of sending the unchanged page files: .html, .js, .css, etc.?
I think you need to understand how JS (or frameworks like React, Vue or Angular) do this. Basically you download the js code on the client, and the js makes all the functionality to update backend and frontend accordingly. In order to connect frontend with backend, use Axios for example.
Can I send the user only the changed, dynamically created html file, and not send again unchanged .js and .css files?
See answer above. Use frameworks like React or Vue, will help you a lot.
Does every user who opens the home page (GET, '/') create a new connection to the server using WebSocket / socket.io?
Depends on what you code. But by default what happens is the user will make a new GET request everytime he accesses your domain, and that's it. (It's not establishing any connection if you don't tell the code to do so).
Hope this helps!! Happy coding!
I'm quite new to expressjs and I'm developing a web application which acts as an API application. There is a react frontend application also. When a button is clicked in the client app it will send an API call to the backend app and will download a file. That scenario is working fine.with the following code.
const file = `${__dirname}/upload-folder/dramaticpenguin.MOV`;
res.download(file); // Set disposition and send it.
});
But now I have a requirement to download multiple files from a button click. How can I do that .
Can someone help me here.
An HTTP response can only have one file. Really, "downloading a file" in HTTP means serving a response with a Content-Disposition: attachment header, to hint to the client that this response should be saved to the filesystem instead of rendered in the browser.
To download multiple files, you want the client code to initiate multiple HTTP requests (probably to different URLs), and the server can respond to each request with a different file. Note that many browsers will refuse to download multiple files in response to a single user action (for fear of flooding the user with unwanted files) or will at least prompt for confirmation before doing so.
If you cannot change your client-side code to make multiple requests, you will need to package your files inside a single file archive.
Context:
I'm making a React website that draws information from the Google Sheets API and formats specific rows into a data visualization. There are columns I don't want to share because of sensitivity of information, and fortunately there are ways to share only specified columns, but that isn't why I'm asking the following:
Problem:
I want to have a Node API that handles requests from a React front-end, but whose code isn't available on the client's browser (for example, in the bundle.js file created during build).
Clarification: I have noticed that when running most Node-React application examples locally and when building them with webpack, you end up with one bundle.js file that contains Node request-handling code being delivered to the browser on page load.
Proposal:
Do I need to deploy two separate apps (one for Node, the other for React), or can I keep them together without the server code being visible to the client?
EDIT POST ANSWER:
you end up with one bundle.js file that contains Node request-handling code being delivered to the browser on page load.
This was untrue. The code I had assumed to be request-handling code was client side request-calling code.
It is already decoupled. There is nothing you need to do.
Note that the security of your node.js server code depends on your server configuration, not node.js. If you access your server via unencrypted file sharing or FTP then your node server code is still not safe.
Even when using encryption, avoid compromised protocols such as SSL or TLSv1.0 (use TLSv1.3 instead for things like FTPS)
You can add a simple authentication system. There are plenty packages out there for Node already, so no need to implement it yourself.
Specifically, this would prevent the backend from sending sensitive data to a unauthorized request.
EDIT: Just for clarification, code run on a Node.js server is not sent out publicly, it will run on your server and send its output to the frontend.
EDIT 2: Looks like I misunderstood your question.
If your code is not decoupled at the moment it will need to be. All code of a React.js project is sent to the browser. Since there is no backend to handle any kind of access logic, any such logic would have to be in the frontend (React.js), where it could easily be circumvented.
I am having a curious problem where a download request from the browser will be passed onto a microservice via the Express layer that serves the frontend. The service queries the db, and generates a csv on the fly and can stream it.
Right now, I am fetching the file to my Express server, and then streaming it to the browser.
Is there a way I can directly pipe the file to the browser without exposing the micro-service endpoint ?