Next Release Preview

Please make a pull request to describe any changes you wish to highlight in the next release of Spinnaker. These notes will be prepended to the release changelog.

Coming Soon in Release 1.30

Spring Boot 2.4

Spinnaker 1.30 uses Spring Boot 2.4, where 1.29 uses Spring Boot 2.3. Spring Boot 2.4 considers session data cached by Spring Boot 2.3 invalid. Therefore, users with cached sessions will be unable to log in until the invalid information is removed from the cache. Open browser windows to Spinnaker are unresponsive after the deployment until they’re reloaded. Executing:

$ redis-cli keys "spring:session*" | xargs redis-cli del

on Gate’s redis instance removes the cached session information.

doNotEval SpEL helper

The doNotEval SpEL helper makes it possible to skip SpEL evaluation in other SpEL helpers e.g. toJson.

For example, if the evaluation context is defined only fileMap object:

Map<String, Object> fileMap = Collections.singletonMap("owner", "managed-by-${team}");

An exception will be thrown in attempt to get JSON because of fileMap contains SpEL inside.


In the given case fileMap contains SpEL for another tool e.g. Terraform. Use doNotEval to let Spinnaker know that this SpEL should be evaluated by a different tool. No exceptions are thrown this way.


Use a feature flag to enable.

# orca-local.yml

    enabled: true

Artifact handling

Changes to the way artifact constraints on triggers work

If you have a pipeline with multiple triggers using different artifact constraints/expected artifacts, these have for a while been evaluated in an unexpected matter.

Artifact constraints

In this example, even though each trigger has its own artifact(s) defined, when one of the artifacts is present, all of the defined artifact constraints on all triggers are evaluated. If any of them are missing, the pipeline will not trigger. This is fixed in Spinnaker 1.30 to only consider the artifacts that are defined on the triggered trigger (there’s an exeption for artifacts that have defined a default artifact or use prior execution; they will always be considered). If you’ve relied on this bug, you’ll need to add manually add all the artifact constraints to all triggers to replicate the previous behavior.

Binding inline artifacts

When defining an inline artifact, e.g. a deploy manifest, Spinnaker will now resolve it in the same way as it does with artifacts that are defined on triggers. Before you could only use SPeL to resolve version numbers etc. ( #4397 )

Inline artifact

Azure image baking improvements


  • Select an Azure account and bake per account. bake_azure_account_dropdown.png
  • You can bake an image using a Managed Image from your Azure account as a base image in the Bake stage. bake_azure_managed_image_stage.png
  • You can bake an image using a Custom Image as a base by specifying the Publisher, Offer, and SKU. bake_azure_custom_image_stage.png


  • Replace deprecated Azure SDK with the new Azure SDK
  • Add a new AzureManagedImageCachingAgent which is caching managed images from the specified region and resource group
  • Update AzureVMImageLookupController to return the newly cached managed images when managedImages query parameter is set to true


Upgraded CentOS base image to version 7.5.

Addition of includeEvents parameter

includeEvents has been added as a configurable parameter in orca that enables users to fetch Kubernetes Events from clouddriver. This parameter defaults to false and is supported by DeployManifestStage, ScaleManifestStage, DisableManifestStage, EnableManifestStage, PatchManifestStage, ResumeRolloutManifestStage and UndoRolloutManifestStage stages.

Once the parameter has been configured and the value is set to true then orca retrieves the events back from clouddriver. Be aware that the size of the pipeline execution context may (significantly) increase by including events.

Kubectl output

With and, it’s now possible to include kubectl output in the pipeline execution context, by setting kubernetes.jobExecutor.persistTaskOutput to true and then setting kubernetes.jobExecutor.enableTaskOutputForAllAccounts to true to include it for all accounts, or setting the debug property to true for individual accounts.


Users coming to Spinnaker are now more familiar with Blue/Green industry terminology than the Netflix specific phrasing Red/Black

  • The Red/Black rollout strategy is marked as deprecated in the UI. Change done here
  • A Blue/Green rollout strategy is added that is functionally equivalent to the Red/Black rollout strategy. Changes done in clouddriver and orca
  • Introduce a pipeline validator in Front50 that validates if red/black is used when creating/updating a Kubernetes pipeline. For now, it is logging a warning, but it will fail when we remove the red/black Kubernetes traffic management strategy. Change done here

Gate expands support for adding request headers to the response header. Previously limited to X-SPINNAKER-REQUEST-ID, it’s now possible to specify any fields with a X-SPINNAKER prefix via the new interceptors.responseHeader.fields configuration property. The default value is X-SPINNAKER-REQUEST-ID to preserve the previous functionality.




To make the dynamic timeout available, you need to enable the feature flag in Orca and Deck.

Orca overrides the default value rollback timeout - 5min - with a UI input from the user.


    enabled: true