Operations

Run myriarchd in production: installation with Helm, authentication, the settings that matter for security, and troubleshooting.

These pages are for the people who run a Myriarch installation. They are written for a reader who already knows the concepts and wants to operate the system reliably.

PageWhat it covers
Install with HelmThe server on Kubernetes with the Helm chart, from the repository or as a signed OCI chart: image, database, first sign-in, GitOps repository, access for agents and enrollment tokens.
AuthenticationHow people sign in (OpenID Connect, the bootstrap admin) and how agents enroll, with recipes for common identity providers.
High availabilityThe leader failover after a hard failure and the setting that bounds it, and how to size shutdownTimeout and terminationGracePeriodSeconds.
Upgrade from v0.1.0 to v0.2.0Migrations and the rolling update, new and changed settings, the Helm chart, the installers and the deprecated Kubernetes collector files.
Support levels and tested platformsThe support level of every component and the host and Kubernetes platforms that the automated tests run on.
Install the collector with HelmCollectors on Kubernetes from the collector chart, with the unchanged upstream images.
Production checklistThe settings of myriarchd serve that matter for security, and how to check a configuration without starting the server.
TroubleshootingSymptoms, causes and fixes, and the endpoints that help.

Guides for upgrades in general, backup and restore, monitoring, air-gapped installation and an artifact mirror are planned for v0.3.0. Until then, the architecture describes the data model and its retention, the artifact cache and the metrics and health endpoints of the server.

Related: the security model explains the controls behind these settings, the Helm chart README covers Kubernetes, and deploy/README.md lists the deployment artifacts and the host installation.


Authentication

How people and agents authenticate to myriarchd: login paths, the bootstrap admin, OpenID Connect recipes and declarative enrollment tokens.

Production checklist

The settings of myriarchd serve that matter for security, what the server refuses at startup, and how to check a configuration without starting it.

Upgrade from v0.1.0 to v0.2.0

Upgrade notes for v0.2.0: migrations and the rolling update, new and changed settings, the Helm chart, the installers and the deprecated collector files.

Support levels and tested platforms

The support level of every component in v0.2.0 and the platforms of the automated tests: Linux and Windows hosts, and two Kubernetes versions on kind.

High availability

How long a failover of the leader takes after a hard failure, the setting that bounds it, and how to size shutdownTimeout and terminationGracePeriodSeconds.

Troubleshooting

Symptoms, causes and fixes for the local system and for installations, plus the endpoints that help to diagnose them.

Install with Helm

Install the Myriarch server on Kubernetes with the Helm chart: image, database, first sign-in, GitOps repository, access for agents and enrollment tokens.

Install the collector with Helm

Run OpenTelemetry Collectors that Myriarch manages on Kubernetes and OpenShift with the Helm chart myriarch-collector and unchanged upstream images.

Upgrade myriarch

Upgrade the server, the database schema and the updaters without downtime.

Backup and restore

Back up and restore the PostgreSQL database and the GitOps repository.

Air-gapped installation

Run myriarch without internet access: images, the artifact cache and collector packages.

Artifact mirror

Serve collector packages from your own mirror instead of the upstream release pages.

Monitoring

Monitor the myriarch server with its metrics, health endpoints and alerts.