Skip to content
This repository has been archived by the owner on Oct 23, 2024. It is now read-only.

Releases: d2iq-archive/dcos-elastic-service

Elastic 3.1.2-7.6.0

14 Apr 08:41
d0d7538
Compare
Choose a tag to compare

Release notes for Elastic Service version 3.1.2-7.6.0

Bug Fixes

  • Fixed missing exporter task metrics with virtual network enabled. (#76)

Elastic 3.1.1-7.6.0

21 Mar 21:39
57af173
Compare
Choose a tag to compare

Release notes for Elastic Service version 3.1.1-7.6.0

Upgrades

  • Upgraded Elasticsearch and Kibana to 7.6.0
  • Upgraded Elasticsearch OpenJDK version from 11.0.2 to 13.0.2.
  • Upgraded Support Diagnostics version from 7.1.1 to 7.1.5.

Elastic 3.1.0-7.4.1

12 Feb 14:14
327e43d
Compare
Choose a tag to compare

Release notes for Elastic Service version 3.1.0-7.4.1

Upgrades

  • Upgraded Elasticsearch and Kibana to 7.4.1
  • Upgraded SDK version from 0.56.3 to 0.57.3. (#70). The following SDK change logs contain more information:

Bug Fixes

  • Fixed Kibana Dashboard issue. (#72)

Elastic 3.0.0-7.3.2

16 Oct 21:27
e7df9b4
Compare
Choose a tag to compare

Release notes for Elastic Service version 3.0.0-7.3.2

Upgrades

  • Upgraded Elasticsearch and Kibana to 7.3.2. (#45 #47)

  • Upgraded SDK version from 0.56.2 to 0.57.1. (#51)

    Please take a look at the following SDK change logs for more information:

New Features

  • Added Prometheus exporter for metrics. (#41)

    New settings are:

    • prometheus_exporter.prometheus_exporter_enabled (boolean, default true)
    • prometheus_exporter.configuration_options (string, default --es.all --es.indices --es.shards)
    • prometheus_exporter.cpus (number, default 1.0)
    • prometheus_exporter.mem (integer, default 128)
    • prometheus_exporter.disk (integer, default 256)
    • prometheus_exporter.disk_type (string, default ROOT)
    • prometheus_exporter.volume_profile (string)
    • prometheus_exporter.readiness_check.interval (integer, default 5, minimum 5)
    • prometheus_exporter.readiness_check.delay (integer, default 0, minimum 0)
    • prometheus_exporter.readiness_check.timeout (integer, default 10, minimum 10)
  • Added Alerts based on Prometheus metrics. (#50 #51)

  • Added Elasticsearch Clusters Overview dashboards based on Prometheus metrics. (#49)

    Check Elasticsearch Metrics Configuration Documentation

  • Exposed xpack.security.transport.ssl.client_authentication xpack.security.http.ssl.client_authentication xpack.security.transport.ssl.verification_mode (#46)

    New settings are:

    • xpack_security_transport_ssl_client_authentication (string, default optional)
    • xpack_security_transport_ssl_verification_mode (string, default full)
    • xpack_security_http_ssl_client_authentication (string, default none)

Bug Fixes

  • Fixed Kibana foldered service URL (<cluster-url>/service/folders/kibana) for DC/OS EE only. (#38 #39)

Improvements

Elastic 2.7.0-6.8.1

20 Aug 10:27
0b9e03c
Compare
Choose a tag to compare

Upgrades

  • Upgraded Elasticsearch and Kibana to 6.8.1. This Elasticsearch version requires that nodes communicate through TLS if X-Pack Security is enabled. We made it so that when elasticsearch.xpack_security_enabled is set to true the scheduler will provision TLS artifacts and configure nodes to only accept TLS connections for both transport and HTTP mechanisms, similar to what one would get by setting service.security.transport_encryption.enabled to true. (#12)

  • Upgraded SDK version from 0.55.4 to 0.56.2.

    Please take a look at the following SDK change logs for more information:

    Changes: (#34)

  • Replaced Oracle JRE (jre-8u192) with newer OpenJDK (openjdk-jre-8u212b03). (#3060)

New Features

  • Added support for Elasticsearch plugin installation via proxy. New settings are:

    • elasticsearch.plugin_http_proxy_host
    • elasticsearch.plugin_http_proxy_port
    • elasticsearch.plugin_https_proxy_host
    • elasticsearch.plugin_https_proxy_port

    Changes: (#4)

  • Added custom domain support for Elastic. New setting is service.security.custom_domain. This custom domain is used in place of the default autoip.dcos.thisdcos.directory. This can be used to expose the service securely outside of the cluster, but requires setting up external DNS. Check this documentation page for more information. (#9)

  • Made it possible to scale down the amount of coordinator nodes. (#15)

  • Added custom YAML configuration support to Kibana. New setting is: kibana.custom_kibana_yml (#14)

  • Added plugin installation support for Kibana. New setting is: kibana.plugins (#16)

  • Added volume profile support. New settings are:

    • master_nodes.volume_profile
    • data_nodes.volume_profile
    • ingest_nodes.volume_profile
    • coordinator_nodes.volume_profile

    Check the Storage Service documentation for more details. (#18)

  • Added placement constraints capability to Kibana. New setting is kibana.placement (array of placement constraint arrays). Check the Marathon placement constraints documentation for more details. (#27)

  • Added virtual network (CNI) support to Kibana. New settings are:

    • virtual_network_enabled (boolean)
    • virtual_network_name (string, default dcos)
    • virtual_network_plugin_labels (array of objects with key and value keys)

    Changes: (#28)

  • Added scheduler-plan based health checks. (#3069)

Bug Fixes

  • Fixed Kibana service URL (<cluster-url>/service/kibana). This makes Kibana's service URL work out of the box even when X-Pack Security is enabled. This was a limitation in the previous version, where we required Kibana to be exposed with EdgeLB. Which is still possible, but not required anymore. (#3036)
  • Fixed issue where setting http.cors.allow-origin to * would cause the deploy to fail to due the Elasticsearch configuration file be compiled with incorrectly escaped content. (#23)
  • Fixed the duplicate Xmx and Xms JVM flags in the command to start the Elasticsearch process. This didn't actually cause any issues because the JVM only considers the last ones, which were the ones we were configuring, but inspecting the command with ps aux would show both, which could be confusing to operators. (#2975)

Improvements

  • Moved the service source code to its own repository: mesosphere/dcos-elastic-service. (#3107)
  • Made most of the Elasticsearch settings be configurable from the service configuration file. (#20)
  • Exposed xpack.monitoring.exporters.default_exporter.use_ingest setting and defaulted it to false to avoid polluting the Elasticsearch node logs with harmless warnings. (#26)
  • Added MARATHON_SINGLE_INSTANCE_APP and DCOS_PACKAGE_FRAMEWORK_NAME to Kibana's Marathon app labels. (#3056)
  • Made search_remote_connect default to false. (#2999)