Heroku | Cannot GET - node.js

Question

I deployed an API to heroku. One of its functions is file uploading. I made a static folder called "uploads" and upload files using multer package.
app.use('/api/uploads', uploadRouter);
app.use('/uploads', express.static(path.join(__dirname, '/uploads')));
Above code snippet comes from server.js. Below is the router and multer setup.
const storage = multer.diskStorage({
destination(req, file, cb) {
cb(null, 'uploads/');
},
filename(req, file, cb) {
cb(null, `${Date.now()}.jpg`);
},
});
const upload = multer({ storage });
uploadRouter.post('/', upload.single('image'), (req, res) => {
res.send(`https://app.herokuapp.com/${req.file.path}`);
});
uploadRouter.post('/imgs', upload.array('images'), (req, res) => {
try{
let filesArray = [];
req.files.forEach(file => {
filesArray.push(`https://app.herokuapp.com/${file.path}`)
})
res.status(201).send(filesArray);
}catch(error) {
res.status(400).send(error.message);
}
});
Something tells me that my approach is far from an ideal one. However, it used to work just perfect until I introduced multiple file upload. Now, images after being uploaded stay there for a little bit (I can access them via links like https://app.herokuapp.com/uploads/1623782012131.jpg) and disappear over few minutes. Then, these URLs throw the following:
Cannot GET /uploads/1623782012131.jpg
The thing is that my previous images that I uploaded before introducing multiple file upload are still accessible. I thought it had to do with heroku file size limitations but it was not the case since I've only used about 50mb out of 500mb. What might be the reason?

Answer 1

Heroku's filesystem is ephemeral/short lived. This means that any images you save on the disk will disappear after you restart or deploy your app. You can't rely on Heroku's filesystem to store images that need to persist for a longer time. Read more here: https://help.heroku.com/K1PPS2WM/why-are-my-file-uploads-missing-deleted
For a reliable image storage, have a look at cloud storage solutions such as Cloudinary or AWS S3.