v6.0.0-beta3

v6.0.0-beta3

v6.0.0-beta

Beta version - not recommended in production

v6.0.0-beta9

v6.0.0-beta9

v5.1.3

This release includes docker images fixes for release v5.1.2 along with following fixes.
Fixes:

• Argus now can evaluate discovery filtering expressions during periodic update of devices. 
• Argus brings graceful and periodic reconciliation of pod devices when there is mismatch in IP addresses.

To enable graceful reconciliation user needs to add "resync.pods = true" property on kubernetes cluster level on LM portal and restart the argus pod.

Upgrade Steps

• Helm charts version for the new release argus is 0.18.0.
  • Run helm repo update followed by upgrade argus using command:
    • helm upgrade --reuse-values -f argus-config.yaml argus logicmonitor/argus
• Recreate Argus pod if it doesn't recreate automatically. Helm does not recreate pods if there is no change in definitions.

v5.1.2

Fixes:

• Argus now can evaluate discovery filtering expressions during periodic update of devices. 
• Argus brings graceful and periodic reconciliation of pod devices when there is mismatch in IP addresses.

To enable graceful reconciliation user needs to add "resync.pods = true" property on kubernetes cluster level on LM portal and restart the argus pod.

Upgrade Steps

• Helm charts version for the new release argus is 0.18.0.
  • Run helm repo update and followed by upgrade argus using command
    • helm upgrade --reuse-values --set imageTag="v5.1.2" -f argus-config.yaml argus logicmonitor/argus
• Recreate Argus pod if it doesn't recreate automatically. Helm does not recreate pods if there is no change in definitions.

v5.1.1

Release 5.1.1

Fixes

Improved retrieving of device group by using its parentId, so even though multiple device groups with same name exists, argus makes sure to pick the correct groups.
Argus continues add device group even though any of the device group add API call fails.
Argus checks for empty device name before sending update call to santaba.

Upgrade Steps

• Upgrade helm charts to the new release (helm repo update) - argus chart version 0.18.0.
  • Run helm repo update and followed by upgrade argus and collectorset-controller using commands
    • helm upgrade -f argus-config.yaml argus logicmonitor/argus
• Recreate Argus pod if they don't recreate automatically. Helm does not recreate pods if there is no change in definitions.

v5.1.0

Release Documents

Release 5.1.0

What's New

• With this release, user can set the custom properties on cluster and device group levels for enabling events and log collection.
• Argus has added ability to automatically delete Kubernetes devices from "_deleted" device group based on configurable interval.
• Added support for Kubernetes version 1.20 for generating self-links for resources.
  ​

Improvements

• User can configure time intervals for Periodic discovery, Periodic delete and Device cache sync in argus-config.yaml helm configuration.
• Added installed Argus version details along with Argus version installation history properties on cluster level for troubleshooting purpose.
• Improved device cache to maintain the device naming to its consistency.
• Enhance Argus to discover headless services and HPA resources.
  ​

Fixes:

• Fixed a bug where Argus was not adding deleted devices to "_deleted" group even though "deleteDevices" configuration was set as false.
  ​

Custom properties for Kubernetes logs and event collection:

• For collecting Kubernetes logs and events, Argus has added ability to provide custom properties in helm configuration. User can add custom properties in helm charts and Argus will add those properties to respective device groups.
• For events collection, lmlogs.k8sevents.enable has been added at cluster level which is applicable for all the device groups. Default is set as false.
• For logs collection, lmlogs.k8spodlog.enable has been added at pods level. Default is set as false.

To enable events or logs collection, user has to set those properties as True.
​

device_group_props:
  cluster:
    # To enable events collection
    - name: "lmlogs.k8sevents.enable"
      value: "false"
  pods:
    # To enable log collection for all pods
    - name: "lmlogs.k8spodlog.enable"
      value: "false"
  services: []
  deployments: []
  nodes: []
  etcd: []
  hpas: []

User can modify these properties from LM Portal or through helm configuration. On Argus restart, property value would be reset to value set in helm config.
​

Auto delete Kubernetes devices from LMPortal based on configured duration:

• Argus moves terminated resources to _deleted dynamic device group if DeleteDevices parameter is set as false.
• Now user can specify the retention period for these devices using property kubernetes.resourcedeleteafter = P1DT0H0M0S.
• Default value for this property is 1 day, that means device will be permanently deleted from LM portal after 1 day and Maximum possible value is 10days which is configured in GCC kubernetes.autoDeleteTask.maxDeleteAfterDuration = P10DT0H0M0S.
• Currently this property can be found at cluster device group, though user can override this for other device groups.
  ​

Device_group_props:
  cluster:
    # To delete resources from the portal after specified time
    - name: "kubernetes.resourcedeleteafterduration"
      value: "P1DT0H0M0S"
      override: false

​

Externalised Time intervals for periodic sync, periodic delete and device cache:

• User can configure time intervals for Periodic discovery, Periodic delete and Device cache sync in argus-config.yaml helm configuration.
• Default values are set to 30m, 30m, 5m and minimum possible values are 10m, 10m, 5m for periodic sync, periodic delete and device cache sync respectively.
  ​

app_intervals:
  periodic_sync_interval: "30m"
  periodic_delete_interval: "30m"
  cache_sync_interval: "5m"

​

If user specifies the time intervals less than minimum possible values then default value would be considered and respective log would be added in Argus.
Please provide valid value for periodic_sync_interval. Since invalid value is configured, forcefully setting it to default 30m.

If user enters invalid duration format (apart from what is mentioned in https://golang.org/pkg/time/#ParseDuration )in argus-config.yaml helm configuration, then application would be forcefully stopped and respective error would be logged. Until user enters correct format, Argus wont start.
​​

Upgrade Steps

• Remove old LogicMonitor helm chart repo - helm repo remove logicmonitor
• Add new LogicMonitor helm chart repo - helm repo add logicmonitor https://logicmonitor.github.io/k8s-helm-charts
• Upgrade helm charts to the new release (helm repo update) - argus chart version 0.17.0 and collectorset-controller chart version 0.11.0
  • Run helm repo update and followed by upgrade argus and collectorset-controller using commands
    • helm upgrade -f argus-config.yaml argus logicmonitor/argus
    • helm upgrade -f collectorset-controller-configuration.yaml collectorset-controller logicmonitor/collectorset-controller
• Recreate Argus, Collectorset-controller pods if they don't recreate automatically. Helm does not recreate pods if there is no change in definitions.

v5.0.1

Release 5.0.1

Fixes

• Conflicting devices among multiple clusters added to same Logicmonitor portal can now be added to monitor

Upgrade Steps

• Recreate the Argus pod

v5.0.0

Release 5.0.0

What's New

• With this release, Argus will discover HorizontalPodAutoscaler resources into Logicmonitor to monitor them.

Improvements

• Argus brought consistency and uniqueness among device names. See more on Device Naming strategy
• Updated Github hosted Argus documentation
• Adding support to monitor pods deployed on Fargate worker nodes.
• Adding permission to /metrics endpoint to monitor API server.
• User may add custom labels and annotations on Argus and Collector controller generated Kubernetes resources

HorizontalPodAutoscaler Monitoring

With this release, Argus have added capability to monitor HorizotalPodAutoscalers - hereafter referred as HPA.

Monitoring

Argus adds HPA as a device in Logicmonitor. Mainly, The Logicmonitor provided Kubernetes_HPA datasource monitors following aspects of it:

• HPA Health: Ensures HPA is able to address the target resource under scale, HPA is able to retrieve desired metrics from metric adapters, HPA is able to scale up and scale down replicas to desired replicas as suggested by HPA's autoscale algorithm.
• HPA autoscale fluctuations: Datapoints collected as current_replicas and desired_replicas generates the line graph. It helps to analyse the trend of application autoscale behaviour.
• Resource metrics: Datapoints for resource metrics collects out of the box.
• Other 3 categories of metrics in HPA are - Pods, Custom, and External metrics: For the sake of an example, Datasource groovy script has given datapoints each category. User have to modify groovy script to collect metrics for any metrics of these categories.

Device Naming Strategy

Adding plain names as that of kubernetes resource into Logicmonitor leads to conflicts as per the following Kubernetes characteristics:

• Same Kubernetes resource name can exist in more than one namespaces
• Same Kubernetes resource name can exist for different types in same namespaces
• Same Kubernetes resource name can exist in two different clusters. This scenario is very rare depending on existence of identical mirrored clusters viz. Production cluster and Production mirror environment for staging could exist.
  Considering the above possibilities of occurring device name conflicts. Argus is now changed to generate unique device names.

Device name generation

Helm chart configuration parameters fullDisplayNameIncludeNamespace and fullDisplayNameIncludeClusterName takes part in device name generation.
Following are ways to define device name generation strategy:

• fullDisplayNameIncludeNamespace and fullDisplayNameIncludeClusterName both set to false (default): [Resource Name]-[Resource Type] Ex: application1-12345-pod, application1-12345-svc
• fullDisplayNameIncludeNamespace set to true and fullDisplayNameIncludeClusterName to false: [Resource Name]-[Resource Type]-[Namespace] ex: kube-apiserver-123-pod-kube-system
• fullDisplayNameIncludeNamespace set to true and fullDisplayNameIncludeClusterName to true: [Resource Name]-[Resource Type]-[Namespace]-[Cluster Name] ex: kube-apiserver-123-pod-kube-system-prod-cluster. Cluster Name is referred to friendly cluster name given at the time of adding cluster into Logicmonitor.
• fullDisplayNameIncludeNamespace set to false and fullDisplayNameIncludeClusterName to true: Here fullDisplayNameIncludeClusterName supersedes the fullDisplayNameIncludeNamespace hence result would be same as both params true.

How Argus behaves with conflicts when fullDisplayNameIncludeNamespace or fullDisplayNameIncludeClusterName is set to false

On occurrence of conflicts with other devices, Argus adds the new conflicting device with full unique name as [Resource Name]-[Resource Type]-[Namespace]-[Cluster Name] and ensures that device remains under monitoring until user notices the device name conflicts and resolves them by opting in for unique name.
Argus also adds the device into _conflicts device group while handling conflicts to nag the user about conflict occurrences. When user notices conflicts occurrences and opt in to use unique name to resolve them, all the resolved devices automatically removes from _conflicts group.

Custom labels and annotations to Argus and Collectorset Controller generated Kubernetes resources

User may provide Custom labels in respective helm chart configuration file. Helm chart configuration parameter labels and annotations holds the key:value maps. And mentioned labels and annotations gets added on all generated resources viz. deployment, pod, service, configmap, serviceaccount, clusterrole, etc.

Pre Upgrade steps

• You need to add Kubernetes_HPA datasource from Logicmonitor exchange beforehand to start monitoring for HorizontalPodAutoscaler devices.

Upgrade Steps

• Upgrade helm charts to the new release (helm repo update) - argus chart version 0.16.0 and collectorset-controller chart version 0.10.0
• Configure the helm configuration parameters fullDisplayNameIncludeNamespace and fullDisplayNameIncludeClusterName according to your needs. Logicmonitor recommends to set both to true which ensures that no device conflicts occur.
• Run helm repo update and followed by upgrade helm upgrade -f argus-config.yaml argus logicmonitor/argus
• Recreate Argus, Collectorset-controller pods if helm upgrade doesn't recreate automatically. Sometimes, Helm does not recreate pods if there is no change in definitions.

Post Upgrade

• With this release, Argus by default adds devices with name as that of resource name, hence ensure that no devices present in _conflicts groups. If conflict groups has devices then make sure to choose appropriate naming strategy as aforementioned in docs.

Known issues

• Bug reference DEVTS-9846: If few devices does not adhere to chosen naming strategy, Argus may not be able retrieve few devices at the time of first resynchronisation to rename existing devices as per chosen naming strategy. Workaround: Restart Argus few times to get it correct. We will update you once we resolve device retrieval bug.

v4.2.0

Release 4.2.0

Improvements

• Enhanced Argus to discover and add devices with same name across different namespaces
• Collecting Kubernetes version and Argus chart version as kind of telemetry on device group property - cluster device group will be used to store this info.
• Periodically synchronising Devices on LogicMonitor platform with its corresponding kubernetes resource. If kubernetes resource no more exists, Argus will remove the device from LogicMonitor platform.
• User may install LM Container solution using Spinnaker pipeline

Fixes

• Fixed bug to not to allow pod device without system.ips property. This property cannot be updated later on so Argus will not add device until the ip would be available for pod.

Upgrade Steps

• Important: While upgrading Argus chart from older versions to 0.14.0 or higher, you need to pass argus configuration file in helm upgrade command. Follow these steps to create configuration file:
  • Download configuration file template from here
  • Get existing values using command helm get values argus and save in backup configuration file.
  • Put all existing values in downloaded configuration file at its appropriate places. And let the remaining values at their default.
• Upgrade helm charts to the new release (helm repo update) - argus chart version 0.15.0 and collectorset-controller chart version 0.9.0
  • Run helm repo update and followed by upgrade helm upgrade --reuse-values -f argus-config.yaml argus logicmonitor/argus
• Recreate Argus, Collectorset-controller pods if they don't recreate automatically. Helm does not recreate pods if there is no change in definitions.