Hi! :) First time here, will keep the intro brief 🙂 We're about to publish a little Druid extension as open source, thinking it might be useful to other people as well. Before we actually publish it though - would be nice if I can get some feedback from the community. You know - it's one thing to use internally a piece of code you've developed, it's completely different when sharing with other people 🙂 Right! So, this extension is basically an InputFormat for OpenTelemetry data - lets you ingest OpenTelemetry into Druid. Over here we find Druid particularly useful for telemetry analytics, and instead of doing complex ingestion specs to process the OpenTelemetry protobufs - we just repurposed a bit of java code that already had. The result is an

otlp

InputFormat extension that flattens and preprocesses OTEL data for easier querying. You can find the pre-release [here.](https://github.com/mishmash-io/opentelemetry-server-embedded/tree/main/druid-otlp-format) I'm pretty sure there are a number of things that can be improved and would be great if I can get some feedback? P.S. Better docs are coming, but will take a bit of time :)

This is cool Lets add it to the apache druid source code itself. druid-contribs-extension would be a good place to start.

That's alright, personally I think a few things are necessary before that: • possibly a more convenient druid table schema - we'll also publish soon some Apache Superset charts and dashboards that we have around, but I'm still not 100% sure the druid table schemas are as easy to use as they can be • There are a couple of other concerns - like, for example - I don't fully understand schema discovery in druid... The OpenTelemetry signals (with some exceptions) are quite good and stable, so, ideally, I would like to make the otlp format work straight away through schema discovery (for ease of use)... but not sure yet we've actually managed that...