CODEWITHSUNDEEP
firebasefirebase-realtime-databasegoogle-cloud-platformgoogle-compute-enginelatency
Alexander Khitev
Alexander Khitev

Reputation: 6851

Firebase access latency

we have an issue with Firebase access latency. We have Node.js application with Firebase SDK. In the code we have next consequence of requests to Firebase: 1. Get keys by GeoFire 2. Then get serveral (from 3 to 100) branches by keys in parallel. 3. Then get same number of branches from another entity by keys in parallel. In Javascript dialect parallel requests looks like this:

const cardsRefs = map(keys, (key) => this.db.ref('cards').child(key));
return Promise.all(map(cardsRefs, (ref) => {
    return ref
        .once('value')
        .then((snapshot) => snapshot.val())
    })
);

That's all, not so big, I think. But it can take a lot of time (up to 5000ms). And we can't be sure about this time, because it can be 700ms sometimes and sometimes it can be much more (up to 5000ms). We expected more predictible behaviour at this point from Firebase. First, we thought that the reason of unstable latency in wrong choosen Google Cloud location (code was ran on Google Compute Engine). We tried from us-west1 an us-central1. Please, lead us to right datacenter location. But then we rewrote code into Google Function and got same result. Please, tell us, how we can get more predictible and sustainable latency on Firebase requests?

Upvotes: 0

Views: 590

Answers (1)

Alexander Khitev
Alexander Khitev

Reputation: 6851

We migrate our functions with the backend to cloud functions and the situation has improved. Each function approximately began to work faster for 1.5 - 2 seconds. At the moment, we are satisfied with this situation.

Upvotes: 1

Related Questions