Getting started

These pages are the tutorials of the documentation. They teach by doing: you follow the steps in order and end with a working system. Three paths lead to it. From source, Myriarch runs on a development machine without a container runtime and without external services: myriarchd dev starts the server with an embedded PostgreSQL, the example GitOps repository and well-known development logins, and myriarch-sim connects simulated collectors to it. With Docker Compose, one file starts the server, PostgreSQL and the simulator from the release images (published from v0.1.0-rc.3 on). On Kubernetes, one script starts a kind cluster with the Helm charts of the server and of the collector (from v0.2.0).

If you want to know first what myriarch is and when it fits, read What is Myriarch? and the comparison. Manage many OpenTelemetry Collectors with OpAMP and Git leads from the tutorials to a fleet of real hosts.

The pages in How-to guides build on this setup. Start with Simulate a fleet to fill the system with agents, then roll out a configuration change or a collector version.

To try the whole system on Kubernetes, with the Helm charts, real server pods and real collectors, follow Kubernetes on kind. The kind probe in hack/kind/README.md builds the same kind of cluster from zero and runs numbered checks against it. All three tutorials use demo policies with a health wait of 30 seconds, so a rollout takes about a minute after you approve its last wave; the pages say so. For a real installation read Install with Helm, Install the collector with Helm and the production checklist.