Azure Kubernetes
Collect Azure Kubernetes metrics
Configuration¶
Install Func¶
It is recommended to enable TrueWatch integration - extension - DataFlux Func (Automata): all prerequisites are automatically installed, please proceed with the script installation.
For self-deploying Func, refer to Self-deploying Func
Install Script¶
Note: Please prepare the required Azure application registration information in advance and assign the
Monitoring Readerrole to the application registration.
To synchronize monitoring data from Azure Kubernetes Service, we install the corresponding collection script: 「Integration (Azure-Kubernetes service collection)」(ID: integration_azure_kubernetes)
After clicking [Install], input the corresponding parameters:
Azure Tenant ID: Tenant IDAzure Client ID: Application registration Client IDAzure Client Secret Value: Client secret value, note it is not the IDSubscriptions: Subscription ID, multiple subscriptions separated by,
Click [Deploy Startup Script], the system will automatically create Startup script sets and configure the corresponding startup scripts.
After enabling, you can see the corresponding automatic trigger configuration in "Manage / Automatic Trigger Configuration". Click [Execute] to immediately execute once without waiting for the scheduled time. After a while, you can check the execution task records and corresponding logs.
Verification¶
- In "Manage / Automatic Trigger Configuration", confirm whether the corresponding task has the corresponding automatic trigger configuration, and check the corresponding task records and logs for any exceptions.
- In TrueWatch, 「Infrastructure - Resource Catalog」, check if there is asset information.
- In TrueWatch, check if there are corresponding monitoring data in "Metrics".
Metrics¶
Collect Azure Kubernetes Service metrics, you can collect more metrics by configuration Microsoft.ContainerService/managedClusters supported metrics
| Metric Name | Description | Unit |
|---|---|---|
apiserver_current_inflight_requests |
Number of in-flight requests | count |
cluster_autoscaler_cluster_safe_to_autoscale |
Cluster health | boolean |
cluster_autoscaler_scale_down_in_cooldown |
Scale down in cooldown | - |
cluster_autoscaler_unneeded_nodes_count |
Unneeded nodes | count |
cluster_autoscaler_unschedulable_pods_count |
Unschedulable pods | count |
kube_node_status_allocatable_cpu_cores |
Total number of available CPU cores in the managed cluster | count |
kube_node_status_allocatable_memory_bytes |
Total amount of available memory in the managed cluster | count |
kube_node_status_condition |
Status of various node conditions | count |
kube_pod_status_phase |
Number of pods by phase | count |
kube_pod_status_ready |
Number of pods in ready state | count |
node_cpu_usage_millicores |
CPU usage (millicores) | % |
node_disk_usage_bytes |
Disk usage in bytes | bytes |
node_disk_usage_percentage |
Disk usage percentage | % |
node_memory_rss_bytes |
Memory RSS in bytes | bytes |
node_memory_rss_percentage |
Memory RSS percentage | % |
node_memory_working_set_bytes |
Memory working set in bytes | bytes |
node_memory_working_set_percentage |
Memory working set percentage | % |
node_network_in_bytes |
Network incoming bytes | bytes |
node_network_out_bytes |
Network outgoing bytes | bytes |
apiserver_flowcontrol_executing_seats_total |
Number of API Priority and Fairness execution seats currently used by the API Server. | count |
apiserver_memory_usage_percentage_maximum |
Maximum API Server Pod memory usage as a percentage of the current limit across instances. | % |
etcd_cpu_usage_percentage_maximum |
Maximum etcd Pod CPU usage as a percentage of the current limit across instances. | % |
etcd_database_usage_percentage_maximum |
The maximum usage of the etcd database for each instance. | % |
etcd_memory_usage_percentage_maximum |
Each instance etcd Pod memory usage accounts for the maximum percentage of the current limit. | % |
scheduler_schedule_attempts_rate_average |
The number of times the scheduler attempts to schedule Pod per second, which can be distinguished by results such as successful scheduling and failure to schedule. | count/s |