Improve observability of Dapr apps with Side-Eye -- a debugger and dynamic observability tool for distributed systems about dapr HOT 2 OPEN

Thank you for taking the time to make the proposal.

A question, how does side-eye benefit the majority of Dapr end users who do not use Go?

from dapr.

First I'll say that we are generally interested in supporting other languages in Side-Eye. Particularly if users emerge that explicitly ask for a language or another.

While Side-Eye is Go-only, I think it should still be quite useful to applications not written in Go. The Dapr runtime itself and the components are written in Go, so providing deep observability into the sidecars and control plane that everybody runs should be a net benefit regardless of what language the app is written in. Besides being able to observe the internals of Dapr, the sidecars seem to know quite a bit about what the apps are doing, and so this info can be very useful to someone debugging. I'm imagining different reports that can be generated exclusively from the sidecars like, for example, reports about in-flight state storage requests, reports on the state of PubSub topics, distributed locks, proxied gRPC calls. Or traces/samples of PubSub messages published or service invocations.

Then, something that we're very keep to explore for Side-Eye in a language-agnostic way is automatically instrumenting applications at the network layer, and extracting telemetry that way. I'm thinking of intercepting and decoding various network protocols (HTTP, Postgres, Kafka...). So, a user with an application written in any language should be able to say "I'm curious what database queries my app is doing; give me a 30s trace from these nodes". This sort of thing is not Dapr-specific, but it should be relevant to the type of apps that I imagine people deploy with Dapr (and perhaps gradually convert to use more and more Dapr components).

Does this make sense?

from dapr.