After a season of battle-tested production use at Sleeper, I’m happy to open source Triton, a Cassandra ORM for Elixir!

Triton makes is easier to work with Cassandra.

Features include:

  • Easily declare keyspaces, tables, and materialized views
  • Automatic compile-time keyspace, table, and materialized view creation
  • Multi-cluster configuration support
  • Compile-time setup hooks for inserting required data
  • Automatic data validation
  • Automatic data transformation

Triton is built on top of Xandra.

Code Examples

Defining a table:

If the table or materialized views do not exist, Triton creates it for you at compile time.

Defining a materialized view:



There’s more examples on the github page.

Why not make Cassandra work with Ecto?

Ecto was primarily designed for relational databases, and has a lot of features that are RDBMS-specific but make no sense in the context of Cassandra data modeling.

We would be constantly trying to fit a square peg into a round hole.

Snowflake ID generation

In addition to Triton, we also open sourced Snowflake, a distributed snowflake ID generator in Elixir.

Happy New Year!