Skip to content

Changelog»

Changes between v2.2.0 and v2.3.0»

Features»

  • Ansible reloaded! We've shipped the following improvements for our Ansible integration:
    • We've added the ability to specify a custom runtime config in YAML format. Click the dropdown button next to the Trigger button and choose Trigger with custom runtime config.
    • The final runtime configuration (including all custom configurations) can be retrieved for all runs. Use the 3 dots in the top right corner of the run details view and choose Runtime config details.
    • In the run changes views, ansible tasks are listed as separate resources.
    • We've added a new Configuration Management tab to the stack view. This view allows you to display the last status of each item in your Ansible inventory (i.e. the outcome of the last run).
    • We've added logs of task executions to the task details tab.
  • Dashboard - Spacelift's new default view is the dashboard, which provides a high-level overview of your stacks, runs, drift detection schedules, as well as some basic metrics (stack failures, run duration etc.).

Changes between v2.1.1 and v2.2.0»

Features»

  • Our bulk actions feature has been reworked, making it simpler to perform actions on multiple items at once.
  • We've added support for OIDC Based API Keys to provide a more secure way of accessing our API without requiring static credentials.
  • We've added the ability to add custom headers to your audit trail webhooks, making it easier than ever to integrate with external systems.
  • We've added an option to config.json to set load_balancer.subnet_placement to either public (default) or private. If the option is omitted, the default remains public.

Fixes»

  • We've added a new check that automatically fails the run if a Kubernetes worker pod exits without properly marking the run as finished or failed.

Changes between v2.1.0 and v2.1.1»

Fixes»

  • Use boostMatches when trying to find a branch in Bitbucket Datacenter. This fixes an issue where we could fail to find the stack's tracked branch if there were too many other branches in the repo with a similar name (for example development, development-123, etc). The exact number of similar branches depends on your Bitbucket configuration, but defaults to 25.
  • Fixed an issue with the Raw Git integration that caused a server does not support exact SHA1 refspec error to be returned when using a Git server that doesn't support cloning with an exact SHA1 refspec.

Changes between v2.0.0 and v2.1.0»

Warning

You must upgrade to v2.0.0 before installing v2.1.0. If you attempt to upgrade an existing installation running a version older than v2.0.0, the installer will report an error and exit without making any changes.

If you are currently on a version older than v2.0.0 and don't have access to the v2.0.0 installer anymore, please reach out to our support team for new download links.

Features»

  • The stack and module settings have been revamped, making it simpler and more efficient to manage your stacks and modules. One of the standout additions is the new Scheduling and Policies tabs, which have moved from the stack settings to the main stack view. Your schedules and policies are now more easily discoverable, and can also be viewed by team members who don’t have access to stack settings.

Fixes»

  • We've fixed an issue with our initial Disaster Recovery support that meant that IoT policies would not be added to your secondary region when resetting an existing worker pool that was created before DR was configured. The impact of this would have been that workers from the affected pools would not be able to connect to your IoT broker after failing over. No user intervention is required to resolve this - the v2.1.0 installation process will fix any affected worker pools.
  • We've fixed an issue affecting Bitbucket Data Center that prevented branches being retrieved correctly if the repository contained more than 25 branches that contained your stack's branch name as part of their name (for example if a stack has a tracked branch called dev, and other branches exist in the repository like development, dev-1, dev-2, etc). This could prevent the stack settings for an affected stack from being saved.

Changes between v1.3.0 and v2.0.0»

Warning

This release of Self-Hosted involves mandatory downtime during the installation process. We expect that downtime to be between 5 and 10 minutes, but it's important that you do not start the installation at a time you cannot afford Spacelift to be unavailable.

v2.0.0 introduces multi-region failover support to Self-Hosted to help as part of a disaster recovery process. One of these changes involves converting the single-region KMS key used to encrypt sensitive data like stack and context secrets to a multi-region key. In order to do that, the installation includes a migration to convert the data from one key to another.

The migration is performed inside a transaction, and in the case of any errors the changes will be rolled back to avoid a situation where data is encrypted using both old and new keys.

Although we have safeguards in place to ensure the migration is successful, we recommend taking a snapshot of your RDS cluster before performing the installation in case anything goes wrong.

Features»

Fixes»

  • Fixed a misconfiguration that was causing the server logs to be filled with messages containing failed to record HTTP transaction.

Changes between v1.2.1 and v1.3.0»

Features»

  • Added ability to view spaces as a list view, you can now switch between Diagram and List view using the toggle in the page header
  • Added Account details drawer (under user menu) with self hosted version, license information and identity provider data.
  • Updated documentation links to use the currently used self hosted version immediately
  • Added list view customization
  • Added stack settings scheduling and policies tabs
  • Added module list and form redesign

Fixes»

  • Fix: Raw Git does not work with terraform modules

Changes between v1.2.0 and v1.2.1»

Fixes»

  • Fix for an issue where a commit to Bitbucket Datacenter could trigger more stacks than necessary
  • Fix for an issue where crashed workers left runs in a hanging state
  • If custom certificates are defined, Spacelift's internal HTTP client will use those for AWS-related requests as well

Changes between v1.1.0-hotfix.1 and v1.2.0»

Features»

Fixes»

  • Fixed a bug where some of the runs weren't scheduled because we attempted to checkout the same license from License Manager at the same time in parallel. Now license checkouts are serialized to avoid this issue.

Changes between v1.1.0 and v1.1.0-hotfix.1»

Fixes»

  • Fixed an issue where license check-out could fail when multiple runs were scheduled at the same time

Changes between v1.0.0 and v1.1.0»

Features»

Fixes»

  • Various backend and frontend fixes and improvements

Changes between v0.0.12 and v1.0.0»

Features»

Fixes»

  • Improved license check-out logic
  • Fix stale logs display for targeted replans
  • Allow to persist roles and collections installed during run initialization for Ansible stacks automatically
  • Various other backend and frontend fixes and improvements

Changes between v0.0.11 and v0.0.12»

Features»

Fixes»

  • Various backend and frontend fixes and improvements

Changes between v0.0.10 and v0.0.11»

Features»

Fixes»

  • Various backend and frontend fixes and improvements

Changes between v0.0.9 and v0.0.10»

Features»

Fixes»

  • Do not re-create SAML certificate during each install

Changes between v0.0.8 and v0.0.9»

Features»

  • Increase worker default disk size to 40GB.
  • Adding support for Terraform versions up to v1.5.7.
  • Update frontend and backend to the latest versions.

Fixes»

  • Enforce bucket policy to prevent objects getting fetched not using HTTPS.
  • Updated no account ID message to indicate that it is caused by missing AWS credentials in the install script.

Changes between v0.0.7 and v0.0.8»

Features»

Removals»

  • Remove the unused ecs-state-handler Lambda.

Fixes»

  • Improve warning message during installation when changeset contains no changes.
  • Fix role assumption and automatic ECR login in GovCloud regions.
  • Don't incorrectly attempt to report errors to Bugsnag in Self-Hosting (errors were never reported, but this could cause some misleading log entries).
  • Fix crash on run startup if the runner image was missing the ps command.
  • Increase default worker pool size to t3.medium.
  • Increase minimum drain instances to 3 to provide more resilience.