I built a SaaS app with per-tenant caches. Initially I used Redis but ran into scale-up issues, so I tried DragonflyDB. It works well in general, but my Lua script use case isn’t supported by default.
The use case is straightforward: each tenant has cached objects like:
`cache:{tenant_id}:{object_id} → cached JSON/doc`
I also maintain a tag index to find all object IDs with a given tag:
`tag:{tenant_id}:{tag} → set of object_ids (tag example: “pricing”, “profile”)`
When a tag changes (say “pricing”), I use a single Lua script to look up all object IDs in the tag set and then delete their cache entries in one atomic operation.
Distributed Lua is cool. Is your implementation similar to DragonflyDB, which doesn’t allow handling undeclared keys in Lua? For example, if I want to generate a new key dynamically inside a script like:
`local queue_key = "queue:user:" .. uid`
How does your system handle such cases?
It's interesting that you have a Cassandra version. I usually think of Cassandra as a standalone database. Why would you layer one database on top of another?
Yeah, I think they were talking about distributed transactions, Redis only support transaction in a single instance, not in a cluster. You can not run Lua across machines.
I took a quick look at the evaluation you did on your blog section, and the numbers look amazing, a little too good to be true to be honest. It must be quite some effort to build this, who are you guys?
It does seem that there are a lot of activities to replace/enhance/improve Redis lately. Dragonfly, Garnet from Microsoft, and the BSD licensed Redis fork Valkey, to name just a few. Now this.