New features, enhancements, and bug fixes

This release includes a number of new features, enhancements, and bug fixes.

Add endpoint for determining supported applications. Add an endpoint listing the applications that the user can access from Okta.

curl -H "Authorization: Bearer $AUTH_TOKEN" http://192.168.40.10.ip.es.io:22400/api/v1/saas/user/applications

Sample response:

{
    "applications": [
        "training",
        "dream-machine"
    ]
}

Alert users about features outside of subscription level. We now alert you if you are using features that are outside of what is included with your subscription level.

New button for creating a monitoring deployment. Add button to create a monitoring deployment on the Logs and Metrics page when only one or no deployment are available.

Default disk quota multipliers (constructor). The constructor service is updated to recognize default disk quota multipliers for each known cluster type.

The constructor now supports three new environment variables:

Make API keys API endpoint public. You can now manage API keys via the keys API endpoint in addition to managing them through the UI.

Add dedicated master nodes when autoscaling a cluster. The autoscaling request handler can now perform post-processing of scaled cluster specification and automatically add dedicated masters, if the number of nodes (any combination of roles, other than dedicated masters) exceeds the threshold. The default number of nodes is six and is defined in the cluster settings.

Add Support portal link. For users with a Support account, the Support page now includes a link to the Elastic Support portal.

Add node_roles support to the EsClusterTopologyElement. Add a new nodeRoles field to the EsClusterTopologyElement for Elasticsearch version 7.10 and later.

Add Workplace Search sync monitor configs. Add Workplace Search content source sync error configurations to the configuration allowlist (whitelist).

Fix staging adminconsole metrics link. We fixed a bug that caused improper ece.cluster filtering on detailed cluster link in the staging adminconsole.

Fix YAML parser in UI. Fixed a bug where attempting to save YAML settings that only include comments resulted in an error message: The YAML settings are invalid, please check your syntax. Saving these YAML settings now works as expected.

JSON Web Tokens no longer change from plan to plan. Fixed an issue where, when log delivery is enabled, a force restart plan caused instances to be created.

Add metadata to IBM default template for default UI select. The default deployment template is selected by default on IBM regions.

Transform SSO and observability plan sections after deployment defaults are generated. The Elastic Cloud Deployments API allows users to omit an Elastic Stack version when creating a new deployment, which enables the system to choose the latest available version. A bug has been fixed where this behavior caused Elastic Cloud to Kibana SSO to not be available when the version was left out.

Update breadcrumbs on update of content. - Breadcrumbs for the ECE SAML, LDAP, and Active Directory pages were not showing up correctly, which is now fixed.

Fix bootloop error message. Fixed a misleading error message that claimed an instance was boot looping even if it may not have been.

Fix issue where APM or AppSearch could get flagged valid for SSO. The Cloud Deployments API had a bug fixed where requests to create or update a deployment can fail in some cases when APM resources share the same ref_id value as Kibana or Enterprise Search. The bug caused the API to consider APM valid for SSO during preliminary logic to build up the API identity.service_provider plan section, but then failed validation when sending the plan to the constructor component.

Fix JSON search slow logs and deprecation logs. Fixed an issue where, when log delivery is enabled, slow logs don’t always get ingested.

Don’t include SSO realm settings when determining if Kibana basic realm should be included. For Cloud SSO, Cloud does not include a basic auth provider (username + password login form) if the system detects that customers have other auth providers configured through custom Kibana settings. There is a bug where, if customers modify the Kibana settings for the auto-generated SAML realm used for Cloud SSO (with the name in the format cloud-saml-kibana-$KIBANA_ID), the basic auth provider is not added. This bug is now fixed, as we only look at customer-provided auth providers and realms while determining if the basic provider should be included.

Skip cross links update on cluster creation. When creating a deployment we resolve all cross-linked resources in the deployment (resources that are linked to the same Elasticsearch) and we populate those links in cluster data when creating each resource. This means that for each resource linked to an Elasticsearch, we add <cluster_type>_cluster_id: <cluster-id>` in each other resource that is linked to the same Elasticsearch.

Fix snapshot suspension. During plan changes, we suspend automatic snapshotting, so that snapshots do not conflict with any of the changes defined in the plan. There was a bug where, in some cases, this suspension was not lifted at the end of the plan, which caused snapshotting to remain suspended. This bug fix ensures that the snapshot suspension is always lifted when a plan is completed.

Correctly update ES security realms when adding an SSO-enabled resource to a deployment. We fixed a bug where single sign-on (SSO) is not enabled when adding a new Kibana instance (7.7.0 and later) or an Enterprise Search instance (version 7.9.2 and later) to a deployment.