Skip to content

Latest commit

 

History

History
333 lines (246 loc) · 11.6 KB

release-notes-8.7.asciidoc

File metadata and controls

333 lines (246 loc) · 11.6 KB

Release notes

This section summarizes the changes in each release.

Also see:

  • {kibana-ref}/release-notes.html[{kib} release notes]

  • {beats-ref}/release-notes.html[{beats} release notes]

{fleet} and {agent} 8.7.1

Review important information about the {fleet} and {agent} 8.7.1 release.

New features

The 8.7.1 release adds the following new and notable features.

{agent}

  • Add support for feature flags, starting with one to toggle fully qualified domain name (FQDN) reporting in events generated by {agent} components, via the host.name field #2218 #2185

Enhancements

{fleet}

  • The agent policy Host name format selector is now enabled by default #154563

{agent}

  • Diagnostics now preserve the Elastic Agent upgrade watcher logs, allowing upgrades that were rolled back to be debugged #2518 #2262

Bug fixes

{fleet}

  • Fixes an issue where the Advanced options toggle in the policy editor always showing #154612

  • Fixes an issue where the warning icon was unable to display in 8.7 #154119

  • Adds updates to output logic #153226

{agent}

  • Fix parsing of paths from the container-paths.yml file used internally when elastic-agent container is run #2340

  • Fix action acknowledgements taking up to 5 minutes. Fixed OSQuery live query results taking up to five minutes to show up in Kibana #2406 #2410

  • Fixes a bug that logging.level settings were not being respected, coming either from Fleet UI or a config file #2456 #2450

  • Fixes a bug that caused an empty proxy from a Fleet managed {agent} policy to override the proxy set by --proxy-url #2468 #2304 #2447

  • Ensure that the /usr/local/bin directory exists on MacOS during {agent} installation #2490 #2487

  • Fixes a bug that caused the lumberjack input type to be missing from the Filebeat filebeat.spec.yaml file, which is required by the barracuda_cloudgen_firewall integration #2511

  • Fixes a bug that prevented sub-directories from being created under the logs/ path in diagnostics ZIP files #2523

  • Make best effort in copying the run directory on upgrades to avoid unnecessary failures. Fixes intermittent upgrade failures when an osquery is running. #2448 #2433

{fleet} and {agent} 8.7.0

Review important information about the {fleet} and {agent} 8.7.0 release.

Breaking changes

Breaking changes can prevent your application from optimal operation and performance. Before you upgrade, review the breaking changes, then mitigate the impact to your application.

Remove the current_upgrades endpoint

Details
The api/fleet/current_upgrades endpoint has been removed. For more information, refer to #147616.

Impact
When you upgrade to 8.7.0, use the /action_status endpoint.

Remove the preconfiguration API route

Details
The /api/fleet/setup/preconfiguration API, which was released as generally available by error, has been removed. For more information, refer to #147199.

Impact
Do not use /api/fleet/setup/preconfiguration. To manage preconfigured agent policies, use kibana.yml. For more information, check Preconfigured settings.

Known issues

Osquery live query results can take up to five minutes to show up in {kib}.

Details
A known issue in {agent} may prevent live query results from being available in the {kib} UI even though the results have been successfully sent to {es}. For more information, refer to #2066.

Impact
Be aware that the live query results shown in {kib} may be delayed by up to 5 minutes.

Adding a {fleet-server} integration to an agent results in panic if the agent was not bootstrapped with a {fleet-server}.

Details

A panic occurs because the {agent} does not have a fleet.server in the fleet.enc configuration file. When this happens, the agent fails with a message like:

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x8 pc=0x557b8eeafc1d]
goroutine 86 [running]:
github.com/elastic/elastic-agent/internal/pkg/agent/application.FleetServerComponentModifier.func1({0xc000652f00, 0xa, 0x10}, 0x557b8fa8eb92?)
...

For more information, refer to #2170.

Impact

To work around this problem, uninstall the {agent} and install it again with {fleet-server} enabled during the bootstrap process.

There is a known issue when upgrading {agent}s to 8.7.0 that are running Osquery.

Details
{agent}s that have the Osquery Manager integration installed can get stuck in an "Updating" state. For more information, refer to #2433.

Impact
Users can do the following work around the issue:

  • Wait for the 8.7.1 release to upgrade {agent}s to the 8.7.x line.

  • Remove the Osquery Manager integration before upgrading. After the {agent} has upgraded to 8.7.0, add the Osquery Manager integration back to the {agent}.

  • If you encounter this issue and {agent}s are stuck in the "Updating" phase, remove the Osquery Manager integration, upgrade the {agent}, and then add it back.

Note
 you may need to use the {agent} upgrade API in this scenario instead of the UI.

New features

The 8.7.0 release adds the following new and notable features.

{fleet}

  • Add getStatusSummary query parameter to GET /api/fleet/agents API #149963

  • Enable diagnostics feature flag and change query for files to use upload_id #149575

  • Add experimental toggles for doc-value-only #149131

  • Display agent metrics, CPU and memory in the agent list table and agent details page #149119

  • Add rollout period to upgrade action #148240

  • Add per-policy inactivity timeout and use runtime fields for agent status #147552

  • Show dataset combo box for input packages #147015

  • Implement a new UI form to support an {agent} shipper in {fleet} #145755

{agent}

  • Add the Entity Analytics {filebeat} input to {agent} #2196

  • Add the ability for the {agent} to receive a diagnostics action #1703 #1883

Enhancements

{agent}

  • Enhance {agent} monitoring configuration to support {filebeat} /inputs endpoint #2171 #33953

  • Render {agent} configuration when running elastic-agent inspect --variables-wait #2297 #2206

Bug fixes

{fleet-server}

  • Accept raw errors as a fallback to detailed error type. This fixes a bug that enrollment or other operations would fail when an error was returned by Elastic in a raw string instead of JSON format. #2079

{agent}

  • Correctly migrate unencrypted {fleet} configuration when upgrading from versions prior to 8.3 #2256 #2249

  • Restore support for memcached/metrics inputs #2298 #2293

  • Fix the log message emitted by the Upgrade Watcher when it detects a crash #2320

  • Fix incorrect, temporary reporting of an agent as unhealthy during installation #2325 #2272

  • Correct the permissions of the state/data/tmp and state/data/logs folders when {agent} is run as a container #2330 #2315

  • Add a timer to periodically check for scheduled actions #2344 #2343

  • Fix a bug that caused {agent} to not start monitoring new Kubernetes pods until it was restarted #2349 #2269

  • Fix possible causes of deadlocks when {agent} shuts down #2352 #2310

  • Fix permission issue on MacOS Ventura and above when enrolling as part of the installation #2314 #2103