Alloy deployment Architecture

Grafana Alloy
1

Hello community,

I am looking for best practises arround deploying Grafana Alloy.
We have some agent running on Windows servers where we collect metrics and logs (log file and windows event logs).
I have tried to forward these to an OpenTelemetry gateway, which on his turn should forward this to a self managed LGTM stack. This does not seem to be an option because the differences in data processing.
I have the option on also running a Alloy gateway instead of the OpenTelemetry?
I am not sure when, or even if i need to deploy a gateway. Shouyd this be looked at the same way as with the OpenTelemetry setup?

thank you in advance!!

2

I can’t speak on OpenTelemetry, I haven’t really used it myself. But Alloy should be compatible with OTLP, but you do want to make sure you are using otlp metrics and/or logs, instead of prometheus or loki.

You will only need an gateway if you can’t get to your LGTM clusters directly. This is usually the cause if you are segregating network traffic between environments.