Maple now automatically links your Infra components in Kubernetes with your actual services

hmm, what is this? currently on railway which has similar ui and this caught my eye

I’d like to use Maple, but against my on prem Clickhouse, not TinyBird (the most expensive compute known to man). Thoughts?

I already support this and got two companies on this! Have some special deals with them, unsure how I would want to price this, since it's a lot of extra work to maintain so not sure i really want people to self serve use it. What do you think?

Definitely willing to pay for it if it’s a good product! The idea would be I give you a connection string to my OTEL collector formatted stuff and that’s it, ideally? Am I able to try this out with the OSS code today?

Just needs a connection to the clickhouse cluster and maple will automatically manage schema changes etc for you and all the tools like the MCP, api and ingestion will work targeting your self managed instance. Ingestion can either be done through the maple ingestion endpoint or you can install the maple collector yourself and directly ingest into the click house instance. Theoretically you could try this out with self hosted OSS but it’s not that straight forward to setup currently

Ah, is there a way to do it in what other SaaS might call “warehouse native” mode? Basically, I don’t do any ingestion through Maple, it’s just read only.

yeah thats also supported

Vibe slop

tthe unsexy part of this is dealing with resources that have no owner labels at all like argo managed workloads, side-effect resources from helm postinstall hooks

That is sadly true 🥲

You made the canvas with xyflow?

Yeah!

xyflow on shadn. cool stuff bro

gimme github repo i wanna see that

the animation is sick :)

Beautiful! 👌🏻

@0x_Fenrir

cyberpunk kubernetes is really awesome

> You think its a nice tool > Open page > "Native to AI." | "or let your AI agent do it." > uhggg.. what?

Curious whether the linking is inferred from labels/service discovery or parsed from manifests. The autodiscovery path is nicer DX but tends to drift when teams rename Helm charts or use ArgoCD overlays — how do you handle re-binding when the underlying mapping changes?

this looks interesting, need to take a look.

Seeing this interface feels like seeing the key holder by the door every single day. It works well, it looks good, but I just don’t want to see it anymore

Pretty

like the design!

The fuck does this even mean?? We’ve had dashboard forever