Productionize Spinnaker

Spinnaker is a large system, made of many microservices, each intended to be scaled, restarted, and configured independently.

Spinnaker is a large system, made of many microservices, each intended to be scaled, restarted, and configured independently. This provides operators a great degree of flexibility, and allows Spinnaker to handle massive scale (1K+ deployments/day, 10K+ managed machines). However, there is no one-size-fits-all approach for configuring Spinnaker; your organization’s usage patterns will need to inform how to prepare your Spinnaker deployment to be used in production.

These pages are intended to provide both an intuition for how Spinnaker works, as well as concrete tips and advice to make Spinnaker more reliable at scale.

Getting started

Before we can confidently productionize an installation of Spinnaker, we need insight into what’s going on. This insight is provided by Spinnaker’s logs and metrics.

  • Logs

    In the Distributed deployments to Kubernetes, logs are forwarded to whatever logging solution you have configured.

    In the LocalDebian deployments, logs are written to /var/log/spinnaker.

  • Metrics

    Follow the monitoring setup instructions to export Spinnaker’s metrics into a metric store of your choice.


Configure Caching
Configure Scaling
Persistence

Last modified May 7, 2021: docs(migration): fix imgs and links (9a18ce6)