Awesome
nessie_cluster
Small but mighty: a library for forming Erlang clusters in Gleam via DNS.
This is a port of the Elixir DNSCluster
library
to Gleam. (Note this project does not have any affiliation with the Elixir library)
Examples
Fly.io
See the example repo for an example Gleam web service with clustering on Fly.io.
Unsupervised
This is NOT recommended. See below for a supervised example.
gleam add nessie_cluster
gleam add gleam_erlang # for sleeping
import gleam/option.{None}
import gleam/erlang/process
import nessie_cluster
pub fn main_simple() {
let cluster: nessie_cluster.DnsCluster = nessie_cluster.with_query(
nessie_cluster.new(),
nessie_cluster.DnsQuery("api.internal")
)
// The process queries DNS periodically, connecting to nodes
let assert Ok(_subject) = nessie_cluster.start_spec(cluster, None)
process.sleep_forever()
}
Supervised
It is strongly recommended to start the process under a supervisor, ensuring it is restarted if it crashes.
This example also sources the DNS name from an environment variable.
gleam add nessie_cluster
gleam add gleam_erlang
gleam add gleam_otp
import gleam/option.{None}
import gleam/erlang/os
import gleam/erlang/process
import gleam/otp/supervisor
import nessie_cluster
pub fn main() {
// Source the DNS name from an env var
let dns_query = case os.get_env("DISCOVERY_DNS_NAME") {
// DNS queries will occur periodically for the given DNS name
Ok(dns_name) -> nessie_cluster.DnsQuery(dns_name)
// ensures DNS queries never occur, e.g. for local development
Error(Nil) -> nessie_cluster.Ignore
}
let cluster: nessie_cluster.DnsCluster =
nessie_cluster.with_query(nessie_cluster.new(), dns_query)
let cluster_worker =
supervisor.worker(fn(_) {
nessie_cluster.start_spec(cluster, None)
})
let children = fn(children) {
children
|> supervisor.add(cluster_worker)
// add other children to your supervisor, e.g. a web server ...
}
let assert Ok(_) = supervisor.start(children)
process.sleep_forever()
}
Further documentation can be found at https://hexdocs.pm/nessie_cluster.
Development
gleam run # Run the project
gleam test # Run the tests
gleam shell # Run an Erlang shell