Released on: 2020-08-07

• Add compliance check command to the DCA CLI (#5930)
• Add clusterchecks rebalance command (#5839)
• Add collection of additional Kubernetes resource types (deployments, replicaSets and services) for Live Containers (#6082, #5999)

• Support "ignore AD tags" parameter for cluster/endpoint checks (#6115)
• Use APIserver connection retrier (#6106)

Released on: 2020-07-20

This version contains the changes released with version 7.21.0 of the core agent. Please refer to the CHANGELOG.

• Add support of DatadogMetric CRD to allow autoscaling based on arbitrary queries (#5384)
• Add Admission Controller to inject Entity ID, standard tags and agent host (useful in serverless environments)

• Add leader_election_is_leader metric to allow label joins (#5819)

Released on: 2020-06-11

This version contains the changes released with version 7.20.0 of the core agent. Please refer to the CHANGELOG.

• Wait for client-go cache to sync for endpoints/services (#5291)
• Consider check failure in advanced rebalancing (#5441)

• Autodiscover standard tags for Cluster and Endpoint Checks (#5241)

• Adds a metric to monitor the advanced dispatching algorithm (#4970)

Released on: 2020-02-11

Minor release on 1.5 branch

• Fix agent commands in DCA (always start listener) (#4870)

Released on: 2020-02-06

Minor release on 1.5 branch

• [DCA] fix cluster-agent flare panic (#4838)
• Remove setcap NET_BIND_SERVICE as we cannot make it work with user namespaces used in the CI (#4846)
• Add service listener in endpoints to watch for newly annotated services (#4816)
• Fix typo (#4831)

Released on: 2020-01-28

This version contains the changes released with version 7.17.0 of the core agent. Please refer to the CHANGELOG.

• Adding logic to show DCA status for clc (#4738)
• Introduce Rate Limiting Stats in the /metrics of the Cluster Agent (#4669)
• MetricServer generates k8s event on HPA

• Add cluster-name tag in host tags (#4558)
• Add read-secret command in cluster-agent to use as secrets backend (#4639)
• Adding logic to show DCA status for clc (#4738)
• Allow dots in cluster names (#4611)
• Check if CheckMetadata exist before iterating over it in cluster agent status page (#4728)
• Grant CAP_NET_BIND_SERVICE capability to the cluster_agent (#4439)
• Ignore invalid cluster names instead of panicking (#4549)
• Fix eventrecorder init (#4732)
• Handle NewHandler failure better in setupClusterCheck (#4447)
• Adding User-Agent to the DCA client
• Filter non-cluster-checks (#4566)

Released on: 2019-11-06

This version contains the changes released with version 6.15.0 of the core agent. Please refer to the CHANGELOG.

• Introducing the Advanced dispatching logic to rebalancing Cluster Level Checks [#4068, #4226, #4344]
• Enable the Endpoint check logic [#3853, #3704]
• HTTP proxy support for the external metrics provider #4191
• Improve External Metrics Provider resiliency [#4285, #3727]
• Revamp the Kubernetes event collection check [#4259, #4346, #4342, #4337, #4314]

• Update Gopkg.lock with new import #3837
• Fix kubernetes_apiserver default config file #3854
• Fix registration of the External Metrics Server's API #4233
• Fixing status of the Cluster Agent if the External Metrics Provider is not enabled #4277
• Fix how the endpoints check source is displayed in agent command outputs #4357
• Fix how we invalidate changed Endpoints config #4363
• Get Cluster Level Checks runner IPs from headers #4386
• Fixing output of agent status #4352

2019-07-09

• Fix Cluster-agent failure with cluster-agent flare command.

2019-06-19

• Fix "Kube Services" service: kube service tags attached to pod are not consistent.

Released on: 2019-05-07

The Datadog Cluster Agent can now auto-discover config templates for kubernetes endpoints checks and expose them to node Agents via its API. This feature is compatible with the version 6.12.0 and up of the Datadog Agent.

Refer to the official documentation to read more about this feature.

2019-05-03

• Fix race condition: immutable MetaBundle stored in DCA cache.

2019-04-30

• Fix race condition in Cluster Agent's API handler.

2019-04-24

• The Cluster Agent can now auto-discover config templates for kubernetes endpoints checks and expose them to node Agents via its API
• Add the config and configcheck command to the cluster agent CLI
• Add the diagnose command to the cluster agent CLI and flare
• Add cluster_checks.extra_tags option to allow users to add tags globally to the cluster level checks.

• Improving Lifecycle of the External Metrics Provider
• Support milliquantities for the External Metrics Provider
• Move some logs from info to debug, in order to generates fewer noisy logs when running correctly.

Released on: 2019-02-25

The Datadog Agent now supports distributing Cluster Level Checks. This feature is compatible with the version 6.9.0 and up of the Datadog Agent.

Refer to the official documentation to read more about this feature.

2019-02-14

• Ensure dangling cluster checks can be re-scheduled

2019-02-12

• Fix re-scheduling of the same clusterchecks config on the same node

2019-02-11

• Sign docker images when pushing to Docker Hub

• Fix configcheck verbose output
• Fix AutoDiscovery rescheduling issue when no template variables
• Remove resolved configs when template are removed
• Support adding/removing the AD annotation to an existing kube service
• Only expose cluster-check prometheus metrics when leading
• Fix support for custom metrics case sensitivity

2019-02-05

The External Metrics Provider is now agnostic of the case, both on the metric name and the labels extracted from HPAs.

• Cluster Agent HPA metrics case support

• Add GetLeaderIP method to LeaderEngine
• Add kube_service config provider
• Allow to set additional Autodiscovery sources by envvars
• Add dispatching metrics in clusterchecks module
• Add a health probe in the ccheck dispatching logic
• Add kube-services AD listener
• Cluster-checks: handle leader election and follower->leader redirection
• Enable clusterchecks in DCA master
• Support /conf.d in cluster-agent image
• Fix clustercheck leader not starting its dispatching logic
• Use the appropriate port when redirecting node-agents to leader
• Cluster-checks: patch configurations on schedule
• Add configcheck/config cmd on the cluster agent
• Add clustercheck info to the cluster-agent's status and flare
• Make error in clusterchecks cmd clear when feature is disabled

2019-01-31

The release of the RC1 was dismissed to embed a fix for the CI runners used to build the image. - Go 1.11.5 compliancy + 1.11.5 for every CI The official release of the Datadog Cluster Agent 1.2.0 starts with the RC2.

The version 1.1.0 of the Cluster Agent introduces new features and enhancements around the External Metrics Provider.

2018-11-21

• Get goautoneg from github
• Fix datadog external metric query when no label is set

2018-11-20

• Migrating back to official custom metrics lib
• Change test to remove flakiness

• Disable cluster checks in cluster-agent 1.1.x
• Allow users to change the custom metric provider port, to run as non-root
• Adding rollup and fix to circumvent time aggregation
• clusterchecks: simple dispatching logic
• Honor external metrics provider settings in cluster-agent status
• Run cluster-agent as non-root, support read-only rootfs
• Only push cluster-agent-dev:master from master

• Fix folder permissions on containerd
• Adding fix for edge case in external metrics
• Fix flare if can't access APIServer
• DCA: fix custom metrics server
• Avoid panicking for missing fields in HPA

Released on: 2018-10-18

The Datadog Cluster Agent is compatible with versions 6.5.1 and up of the Datadog Agent.

• Please refer to the 6.5.0 tag on datadog-agent for the list of changes on the Datadog Agent.

It is only supported in containerized environments.

• Please find the image on our Docker Hub.

2018-10-17

• Expose telemetry metrics with the Open Metrics format instead of expvar

• add mutex logic and safe guards to avoid race condition in the Autoscalers Controller.

2018-10-15

• Leverage diff logic to only update the internal custom metrics store and Config Map with relevant changes.
• Better logging on the Autoscalers Controller

• Make sure only the leader sync Autoscalers.
• Forget keys from the informer's queue to avoid borking the Autoscalers Controller.

2018-10-11

• Support agent and datadog-cluster-agent for the CLI of the Datadog Cluster Agent
• Retrieve hostname in GCE

2018-10-04

• Implement the External Metrics Interface to allow for the Horizontal Pod Autoscalers to be based off of Datadog metrics.
• Use informers to be up to date with the Horizontal Pod Autoscalers object in the cluster.
• Implement the metadata mapper.
• Use informers to be up to date with the Endpoints and Nodes objects in the cluster.
• Serve cluster level metadata on an external endpoint, kube_service tag is available.
• Serve node labels as tags.
• Run the kube_apiserver check to collect events and run a service check against each component of the Control Plane.
• Implements the flare, status and version commands similar to the node agent.