Firestore automatic backup using cloud functions? - firebase

Question

Firebase docs recommends that you deploy an App Engine application to handle automatic firestore exports.
https://firebase.google.com/docs/firestore/solutions/schedule-export
app.js
const axios = require('axios');
const dateformat = require('dateformat');
const express = require('express');
const { google } = require('googleapis');
const app = express();
// Trigger a backup
app.get('/cloud-firestore-export', async (req, res) => {
const auth = await google.auth.getClient({
scopes: ['https://www.googleapis.com/auth/datastore']
});
const accessTokenResponse = await auth.getAccessToken();
const accessToken = accessTokenResponse.token;
const headers = {
'Content-Type': 'application/json',
Authorization: 'Bearer ' + accessToken
};
const outputUriPrefix = req.param('outputUriPrefix');
if (!(outputUriPrefix && outputUriPrefix.indexOf('gs://') == 0)) {
res.status(500).send(`Malformed outputUriPrefix: ${outputUriPrefix}`);
}
// Construct a backup path folder based on the timestamp
const timestamp = dateformat(Date.now(), 'yyyy-mm-dd-HH-MM-ss');
let path = outputUriPrefix;
if (path.endsWith('/')) {
path += timestamp;
} else {
path += '/' + timestamp;
}
const body = {
outputUriPrefix: path
};
// If specified, mark specific collections for backup
const collectionParam = req.param('collections');
if (collectionParam) {
body.collectionIds = collectionParam.split(',');
}
const projectId = process.env.GOOGLE_CLOUD_PROJECT;
const url = `https://firestore.googleapis.com/v1beta1/projects/${projectId}/databases/(default):exportDocuments`;
try {
const response = await axios.post(url, body, { headers: headers });
res
.status(200)
.send(response.data)
.end();
} catch (e) {
if (e.response) {
console.warn(e.response.data);
}
res
.status(500)
.send('Could not start backup: ' + e)
.end();
}
});
// Index page, just to make it easy to see if the app is working.
app.get('/', (req, res) => {
res
.status(200)
.send('[scheduled-backups]: Hello, world!')
.end();
});
// Start the server
const PORT = process.env.PORT || 6060;
app.listen(PORT, () => {
console.log(`App listening on port ${PORT}`);
console.log('Press Ctrl+C to quit.');
});
package.json
{
"name": "solution-scheduled-backups",
"version": "1.0.0",
"description": "Scheduled Cloud Firestore backups via AppEngine cron",
"main": "app.js",
"engines": {
"node": "8.x.x"
},
"scripts": {
"deploy": "gcloud app deploy --quiet app.yaml cron.yaml",
"start": "node app.js"
},
"author": "Google, Inc.",
"license": "Apache-2.0",
"dependencies": {
"axios": "^0.18.0",
"dateformat": "^3.0.3",
"express": "^4.16.4",
"googleapis": "^38.0.0"
},
"devDependencies": {
"prettier": "^1.16.4"
}
}
cron.yaml
cron:
- description: "Daily Cloud Firestore Export"
url: /cloud-firestore-export?outputUriPrefix=gs://BUCKET_NAME[/PATH]&collections=test1,test2
target: cloud-firestore-admin
schedule: every 24 hours
QUESTION
But I was wondering, is it possible to achieve the same thing with an HTTP cloud function and the Cloud Scheduler ?
Is there anything in this App Engine code that I couldn't reproduce or access with an HTTP cloud function? I mean, is the App Engine project really necessary here?
NOTE: This is not an opinion-based question, nor it's too broad. I want to know if I need the App Engine to achieve this behavior or not, and why.
I wouldn't need to set it as an express server, obviously. Just a normal HTTP cloud function, that when called, does the exporting.
I would add something like the following cron job into the Cloud Scheduler:

Answer 1

You will not need App Engine at all if you use Cloud Scheduler to trigger a function that does the backup.