This repository is personal study to understand the observability software stack of microservices.
- First, create a specific namespace for the tools/systems working on the control-plane managements
$ kubectl create namespace observ
The namespace is important for both using the deployment files in this repo and doing service discovery (DNS). If you change this, please change the files accordingly.
- In certain test, you will need to setup a Prometheus server on your local host.
Please refer to here, but be awared of changing the namespace to
observ
Metrics/Traces/Logs are published directly (PUSH) from application. In this mannar, I demonstrate two pipelines:
- Metrics: using Prometheus as the backend, requiring to have OTel collector in the middle stage.
- Traces: using Jaeger as the backend, getting traces straightforward from application side.