The shards that hold the streams, a Slap.Cluster. Each shard allocates stream ids and supervises one process per active stream.

Run one instance of this built-in cluster per VM. It can own many shards and participate in a cluster across VMs. Slap.Streams.Metrics reports node-wide metrics for this instance.

To configure a cluster under your application's name, define a module with use Slap.Streams.Cluster, otp_app: :my_app, supervise that module in place of Slap.Streams.Cluster, and pass cluster: MyApp.StreamsCluster to Streams operations and Slap.Streams.HTTP.Router. Run one Streams cluster per VM; metrics and the router's request-body budget are node-wide.

Configure it like any Slap.Cluster:

config :slap_streams, Slap.Streams.Cluster,
  store: {:url, "s3://streams/db", aws_endpoint: "https://rustfs:9000"},
  shards: 8

Durable appends wait for SlateDB's next WAL flush. This cluster sets flush_interval to 10 ms, the value used by the Durable Streams design; SlateDB's default is 100 ms. Other settings leave the 10 ms default intact. The README describes the latency and object-store request trade-off.

Service settings go in child_options: at start. See child_option_keys/0 for the accepted keys. Values are positive integers in milliseconds for timeouts and intervals, bytes for byte limits, and entries for :deleter_page. :max_fork_copy_bytes and :fork_copy_grace may also be zero.