API Rate Limiting
If your Spinnaker deployment interacts with API clients, the API rate limiter can help keep it reliable during traffic spikes and protect it from rogue clients.
Runbooks
Runbooks on how to configure different options available for Spinnaker.
Artifact Store
Spinnaker copies embedded artifacts into the execution context every time a stage touches them. The artifact store persists them once to external storage and leaves a small reference behind, which can cut execution size dramatically.
Caching: Account Name Based Sharding
How to shard caching agents among the available caching pods based on the account name
Reducing Execution Size with the Entity Store
The entity store extends artifact storage to arbitrary objects in the execution context, most usefully Kubernetes manifests. Manifests are stored once externally and replaced with a small reference, which has cut execution size by up to 85% on large pipelines.
Orca: Reducing Execution Size by Excluding Keys from Stage Outputs
Some Orca tasks copy large objects such as manifests and artifacts into a stage’s outputs, which are then propagated for the rest of the execution. You can configure Orca to exclude those keys and keep executions considerably smaller.
Orca: Quality of Service
Spinnaker includes an optional Quality of Service (QoS) module you can use to manage how many active executions run simlutaneously.
Orca: Zombie Executions
A zombie execution is one whose status is RUNNING, but there are no messages in Orca’s work queue or unacked set.
Sharding Spinnaker
How to shard traffic to different areas of Spinnaker, in case a service doesn’t efficiently serve all queries with a single instance of Orca or Clouddriver.