#360 implemented ingestion and serving support for Cassandra, for the v0.3 series of Feast. It is considered experimental: we the implementers have not run it in production and are in the process of stress testing.
Project Namespacing (#359 and others) and async job management (#361), coming along near the same time, brought significant changes to Feast internals for v0.4. Consequently, forward-porting the Cassandra support is not trivial.
As contributors of the functionality, we are evaluating our timeline for whether we can change course for Feast v0.4+, or will need to go to production on v0.3 and update on a later schedule. In the former case, we will take on this porting task in the next 1-2 months (Jan-Feb), but in the latter case, we would welcome anyone else interested in Cassandra for v0.4+ to run with implementing it—tag @smadarasmi and @ches and we'll be happy to review, or answer questions about the existing implementation; I'm also in the public Feast Slack channel.
We will also most likely wait for dust to settle on decisions like Feature Set Versioning (#386) before we revisit the implementation.
I will update this issue accordingly when our timeline is clear.
#360 implemented ingestion and serving support for Cassandra, for the v0.3 series of Feast. It is considered experimental: we the implementers have not run it in production and are in the process of stress testing.
Project Namespacing (#359 and others) and async job management (#361), coming along near the same time, brought significant changes to Feast internals for v0.4. Consequently, forward-porting the Cassandra support is not trivial.
As contributors of the functionality, we are evaluating our timeline for whether we can change course for Feast v0.4+, or will need to go to production on v0.3 and update on a later schedule. In the former case, we will take on this porting task in the next 1-2 months (Jan-Feb), but in the latter case, we would welcome anyone else interested in Cassandra for v0.4+ to run with implementing it—tag @smadarasmi and @ches and we'll be happy to review, or answer questions about the existing implementation; I'm also in the public Feast Slack channel.
We will also most likely wait for dust to settle on decisions like Feature Set Versioning (#386) before we revisit the implementation.
I will update this issue accordingly when our timeline is clear.