Redis

after it deployed successfully.

But it was like, a Redis and Action Cable don't seem to be connecting properly.

So then it goes and reads the source code for Campfire.

And Claude then was like, oh, they've hard-coded this so that it uses the Redis host name because it's using the Docker container name.

And let me just go in Hatchbox, create these two overrides during deployments and override these two files.

And then that way you can use the environment variable to change the Redis URL to what Hatchbox needs.

It can use SQLite and everything like the ones products use.

And it's wild that it could override it.

And that will sum across every queue connection that it knows about and give you those values back.

This is implemented for the Redis database failover and fake drivers.

Thanks, Jack Bayless, for that one.

Vectors distance queries on MariaDB, they were previously where vectors similar to, and they have been moved to the grammar following the pattern already used for, let's see here.

Spark, uh, you know, Databricks was really successful actually, and Confluent pretty well, uh, being acquired after by IBM.

Uh, Redis, that's another story that actually didn't go really well.

Um-

Yeah

Yeah.

So basically what you're saying is something like caching, like Redis, that's a good candidate for running a Kubernetes stuff like Mongo and your actual, I think you said it quite nicely when you said, well, any critical databases, keep them outside the Kubernetes cluster unless you really want to.

So that's cool.

So I'm going to close this section by asking you to complete the sentence.

You have the right teams, right knowledge.

you can run databases on Kubernetes.

And then in the end, we came to, you know, you mentioned some of the workloads that fit really well to run on like databases on Kubernetes, things like caches that you can run like Redis.

So that's great.

And lastly is caching.

So we've always had a Redis cache.

So the Redis cache, I guess I'll get into that in a sec.

So the issue is that even with PostGIS, some of the complex queries are costly, and so we want to avoid repeating them as often as possible.

So we're taking kind of a three-pronged approach right now.

If your cache is already full, it throws away, for example, the thing that no one's accessed in three weeks and just kind of cycle like that.

So this isn't working perfectly for us yet.

I think originally we had thought that we didn't need to do all of it and started ignoring the Redis caching.