Chris Biscardi

moon indicating dark mode
sun indicating light mode

A first look at FaunaDB

FaunaDB is a geo-replicated database in the style of Spanner and Calvin. It includes a GraphQL integration natively, is multi-tenant by default (using nested databases), and has a powerful query language based in lambda calculus called FQL.

Also check out the Jepsen analysis and the consistency models chart if you’re unfamiliar with consistency models.

Most interestingly is that it’s bonkers easy to set up a new database for a project (ymmv, statment is subject to chance and experience, void where prohibited).

A first document insertion

My use case for fauna is a Golang lambda function that I deploy on Netlify that runs once a day at 11am. I would like to store the Dev Tip that gets tweeted out somewhere so that I can prevent it from being tweeted again within a week or two of the last time it was tweeted out.

First, I created a FaunaTweetResult type complete with fauna tags.

type FaunaTweetResult struct {
DevTipID string `fauna:"devTipId"`
TweetID string `fauna:"tweetId"`
TweetContent string `fauna:"tweetContent"`
TweetCreatedAt string `fauna:"tweetCreatedAt"`
RetweetCount int `fauna:"retweetCount"`
FavoriteCount int `fauna:"favoriteCount"`

Then I instantiated a fake tweet that matches the output of the function I’m using to send the tweet in anaconda

faunaTweet := FaunaTweetResult{
DevTipID: "somethingtip",
TweetID: "some tweet id",
TweetContent: "some tweet",
TweetCreatedAt: "a date",
RetweetCount: 0,
FavoriteCount: 0,

Next comes the token fetch from the environment and client instantiation for fauna. You can create sub-clients using alternate secrets which is interesting but I didn’t have a use case for it yet.

faunaToken, _ := os.LookupEnv("FAUNA_TOKEN")
client := f.NewFaunaClient(faunaToken)

Finally we write some FQL looking code and pass in our faked tweet to insert a new document into our fauna database.

newTweetRecord, faunaErr := client.Query(
// f.Class will be f.Collection soon
f.Obj{"data": faunaTweet},

and that’s it. We were successfully inserting docs into the database in about an hours worth of work, having never worked with the platform before.

It was pretty fun to throw my proverbial hat in the faunadb ring and I’ll be working with it more in the future. My next task is to tackle querying so that I can develop a better mental model for how to architect data for this system.