Now you can override your values.yml at the K8s Apply step in a workflow. You can do it either inline or remote (via Git repositories).
This makes the deployment more flexible as now you can have multiple Apply steps in a workflow and then override the values at runtime as well.
We are supporting the instance sync for SSH deployments (PDC, AWS & Azure) by which the service dashboard will contain the real-time values of instances that are up for these services.
Users now have the ability to resolve tags during execution and won’t need to wait until the end.
Now we are supporting exponential backoff strategy for all the CF operations. With this if there is any failure in a cloud formation step, it won't fail by retry.
This is going to be configurable by setting some account-level defaults.
You can refer to the docs to know more about it here
Earlier for SSH and WinRm deployments, in Select Nodes step user can choose to deploy specific hosts and choose from a list of hosts added previously in Infrastructure Definition.
Now you can specify hosts which are not present in Infrastructure Definition. Doc reference is here
User can also choose hosts which are not defined in Infrastructure Definition for SSH deployments on Start New Deployment screen with the option to target specific hosts enabled.
Doc reference is here
This is behind a feature flag: DEPLOY_TO_INLINE_HOSTS
Now you have an app-level setting by which you can make it mandatory for all the github triggers to contain Webhook secrets or else the invocations will fail. This adds an extra layer of security as now the admins can choose to not allow any triggers which are not containing Webhook secrets.
This is going to be behind the Feature flag - GITHUB_WEBHOOK_AUTHENTICATION
As part of MVP we built the support for having a new option in the failure strategy as Rollback Provisioners after Phases at workflow & step level.
We are extending the same while a manual intervention is waiting for an action & for the manual rollbacks triggered by you.
This is going to be behind the same Feature flag - ROLLBACK_PROVISIONERS_AFTER_PHASES
You can refer to the docs here
We are enhancing our Event rules which were earlier supported for only Pipeline events to now support workflow events (Workflow start, Workflow end, Workflow paused and Workflow continued) as well.
We will send these events in both cases if it’s a direct workflow execution or the workflow is executing as part of the pipeline.
This is going to be behind the same Feature flag - APP_TELEMETRY
You can refer to the docs here
Now you can create/edit Service Now & Artifactory Connectors by skipping the credentials/connectivity verification which will help you setup your connectors even if you don't have delegates who are able to connect to your services or if the services are down for some reason.
We have introduced a checkbox 'Skip Validation' for ServiceNow and Artifactory connectors. Checking this box will allow the user to skip credentials verification while creation/updation to go through.
Default behavior would be followed if the box remains unselected
RBAC in CG now allows customers to define roles that provide permissions to users on all the applications except a few. Till now customers could select all or selective applications when scoping permissions on application but now they will also be able to select applications for exclusion. This helps in catering to scenarios where they want to have strict control over a set of sensitive / critical applications while allowing their teams to manage rest of the applications. This feature automatically includes any new application that is created in Harness in the list of allowed applications but they can be moved to excluded category by explicitly modifying the scope.
In the example below, permissions will apply to all the applications except "K8S Sample App" when "Exclude Selected" option is used
This is feature is behind FF CG_RBAC_EXCLUSION
