The following changes are included in this release.

Featuresedit

Make allocator add direct trust certificates to Elasticsearch. This release of Elasticsearch Service configures Elasticsearch instances to trust certificates uploaded by customers for use with remote connections to Elasticsearch clusters outside of a Cloud managed environment.

New migrate_template endpoint to migrate a deployment to a different deployment template. Adds the migrate_template endpoint to allow migrating a deployment to a different deployment template.

Add type parameter and more validation to DirectTrustRelationship. DirectTrustRelationships now take an optional type parameter which enables use case specific validation and supports a more guided experience in the user interface.

Allow xpack.security.authc.realms.jwt settings. Adds support for the beta JWT realm for Elasticsearch authentication.

Enable using different domain for Microsoft Azure storage. Enables support for the different domain used by blob storage in Azure Govcloud.

Enhancementsedit

Reset the stack version of stateless resources when they are being re-enabled. Explicitly reset the version of a stateless resource (APM, Kibana, Enterprise Search) when re-enabling the resource on the Deployment edit page. Previously, these resources would be re-enabled using the version they were last running on. This blocked further changes when the Elasticsearch resource had been upgraded to different major version (for example 7.x to 8.x).

cloud-security-posture feature flag. Adding xpack.cloudSecurityPosture.enabled: Kibana feature flag for cloud-security-posture.

Replace watcheralert.found.io references with alerts.elastic.co. Changes the email address references for watcher alert emails from watcheralert.found.io to alerts.elastic.co.

Allow monitoring to be shipped to a deployment running the next major version. Allow logs and metrics to be shipped to deployments running the next major version via the UI. For example, a 7.x deployment can now be configured to ship monitoring data to an 8.x deployment.

In runner’s reference.conf, set notification-email-from to Watcher Alert <noreply@alerts.elastic.co> by default. Changes the email address that’s used as sender for the Watcher alert emails from noreply@watcher.elastic-cloud.co to noreply@alerts.elastic.co. The change will apply to newly created deployments and when applying a plan to existing deployments.

Change sender address for Kibana alerts. Changes the email address that’s used as sender for Kibana alert emails from alerts-noreply@elastic.co to noreply@alerts.elastic.co. The change will apply to newly created deployments and when applying a plan to existing deployments.

Don’t show the APM secret token when APM is in managed mode. Stop showing the APM secret token when APM is managed by the Elastic Agent. Instead link through to the Fleet UI containing the secret token.

Adds support for service logs that are sent directly to Firehose. For example Route 53 query logging, Firewall, etc. These logs are configured to send to a given Firehose delivery stream instead of going through CloudWatch log group first.

Output log event looks like this:

{
  "_index": ".ds-logs-aws.route53_resolver-default-2022.04.15-000001",
  "_type": "_doc",
  "_id": "660ddd8aaa",
  "_version": 1,
  "_score": 1,
  "_source": {
    "message": "{\"version\":\"1.100000\",\"account_id\":\"12345\",\"region\":\"us-east-1\",\"vpc_id\":\"vpc-e8722d92\",\"query_timestamp\":\"2022-04-15T22:26:37Z\",\"query_name\":\"tk2-209-14006.vs.sakura.ne.jp.\",\"query_type\":\"A\",\"query_class\":\"IN\",\"rcode\":\"NOERROR\",\"answers\":[{\"Rdata\":\"160.16.66.10\",\"Type\":\"A\",\"Class\":\"IN\"}],\"srcaddr\":\"172.31.84.43\",\"srcport\":\"50231\",\"transport\":\"UDP\",\"srcids\":{\"instance\":\"i-0646c0435554cc5ed\"}}",
    "@timestamp": "2022-04-15T22:27:49Z",
    "cloud.provider": "aws",
    "cloud.account.id": "12345",
    "cloud.region": "us-east-1",
    "aws.firehose.arn": "arn:aws:firehose:us-east-1:12345:deliverystream/route53-resolver-logs",
    "aws.firehose.request_id": "3d919d79-69c0-48b7-aa63-85d2f886a582",
    "aws.kinesis.type": "deliverystream",
    "aws.kinesis.name": "route53-resolver-logs",
    "aws.firehose.parameters.X-Found-Cluster": "ff0b7f06475245199add23784d445d3e",
    "aws.firehose.parameters.es_datastream_name": "logs-aws.route53_resolver-default",
    "data_stream.type": "logs",
    "data_stream.dataset": "aws.route53_resolver",
    "data_stream.namespace": "default"
  }
}

Change Firehose parameter name to es_datastream_name. Changes the Firehose parameter name for users to specify a data stream name/index name to be only es_datastream_name. Using an index is not recommended and should not be exposed in the parameter name.

Adds support for Firehose GZIP content encoding. When configuring Firehose delivery stream, user can choose to uses the content encoding to compress the body of a request before sending the request to the destination.

Don’t check the upgrade rules when re-enabling a terminated stateless resource. Stop enforcing upgrade rules when re-enabling stateless resources. For example, allow APM to be re-enabled on 8.2 if it was disabled on 7.14

Remove CSV maxSizeBytes whitelist cap for 7.15+. Allow xpack.reporting.csv.maxSizeBytes whitelisting to exceed 50MB for cluster versions 7.15+, since it is chunked now.

Allow configuring Crawler content extraction settings. In version 8.3, the crawler in Enterprise Search adds new settings related to content extraction.

Warning when keystore entry is suspected to be invalid. In order to prevent invalid entries to the keystore that might cause all future plan changes to fail, we’ve added a list of well known keys that are expected to work, and validates the user-entered keystore entries against this list.

In case a key is not on the list of well known keys, the UI now issues a warning, provides some extra details on how secure settings work, and also asks the user to confirm that the setting might prevent future plan changes and may put the availability of their clusters at risk.

Reject requests to restart stopped deployment resources not running the same major stack version. Block requests to restart stopped stateless resources when they are not configured with the same major stack version as the Elasticsearch resource.

Beats upgrade to 7.17.1. Filebeat and Metricbeat shipping logs and metrics to the logging-and-metrics cluster in Elastic Cloud Enterprise are upgraded from 7.12 to 7.17.

Set the region when creating a dedicated monitoring deployment. Pre-select the correct region when creating a dedicated monitoring deployment

Fix structured logging exceptions to comply with ECS. Changes the structured logging format to be Elastic Common Schema (ECS) compliant. Changes exception to error and class_name to type.

Bug fixesedit

Change SSO realms to start at order 101, after user settings realms. Elasticsearch Service now uses realm orders above 100 for SSO realms to avoid conflict with any realms defined in user settings.

Show hourly rate instead of total usage in hourly rate tile. Fixes the value shown in the hourly rate tile of the usage page for trial users.

Fixed Summary sidebar. When editing a deployment, the summary sidebar no longer gets squashed when scrolling down the page.

Added more detail when instance overrides are reset. The calculate incremental elasticsearch change step was omitting the reason why the instance resize is done, when dealing with instance capacity overrides. This is fixed now and the reason is included in the detail step view.

Reset instance capacity overrides. Instance capacity overrides were not reset during the next plan change. Now, during the next plan change, the instance capacity overrides are removed and the initial capacity is used.

Fixed overflow problem when displaying summary of resized instances. Fixes the negative numbers that were sometimes shown in the overview of the resize step due to an overflow bug. This occurred when calculating the capacity of an instance in human readable fashion.

Don’t count stopped deployments on the platform region overview page. The healthy/unhealthy deployment counts on the platform overview page now exclude stopped deployments.

Only show the Enterprise Search read-only warning if Enterprise Search is enabled. During upgrades, if Enterprise Search was not enabled, sometimes a warning would be displayed about it being read-only.

Fix Proxy panic on error with empty request body Fixes Proxy panic on retry of url.Error with empty request body.

Deprecationsedit

Change useDiskThreshold to hidden property and don’t parse or serialize. The formerly deprecated use_disk_threshold field in ElasticsearchSystemSettings will no longer be returned in API responses and any value attempted written is ignored.

Update legacy region cipher suite. Updated the legacy region cipher suites to intermediate ciphers + Windows 11 client compatible cipher suites. Read more about this change to the supported ciphers list.

Docsedit

Add docs example for simplified deployment creation. We’ve added a new example to the Elasticsearch Service API docs showing how to create a deployment using default values based on a deployment template specified in the API request.