Kubernetes Monitoring

After you integrate Cisco AppDynamics SaaS and Cisco Cloud Observability, you can troubleshoot the Kubernetes entity related issues with your application. 

This page includes the details required to get a unified view of Cisco AppDynamics SaaS Controller and Cisco Cloud Observability (CCO). This unified view helps troubleshoot Kubernetes-related infrastructure issues within specific applications or tiers monitored in the SaaS Controller.

You use the flowmap view to monitor the application performance in Cisco AppDynamics SaaS, whereas you use the entity details view to monitor the Kubernetes entities in CCO. 

The unified view qualifies the APM flowmap with the correlated Kubernetes workload on the Cisco AppDynamics SaaS Controller UI. The correlated workloads are available when the cluster is monitored by Cisco Cloud Observability Kubernetes collectors. The container details at application and tier level correspond to the respective workloads in Cisco Cloud Observability UI. 

The application owners, or DevOps can use this capability to identify root cause of issues in the applications or tiers by correlating them with the Kubernetes infrastructure layer, which reduces mean time to identify (MTTI) and the mean time to resolve (MTTR) issues.

Before You Begin

Ensure the following requirements are met before you can view the Kubernetes entity details in Cisco Cloud Observability:

• The Controller version >= 24.2.0.
• The licenses are available for both Controller and Observability Platform tenants.
• You are using Cisco AppDynamics SaaS setup.
• The application uses Java or .NET and is instrumented with Java Agent or .NET Agent respectively.
  See Java Agent, .NET Agent, or Auto Instrument Applications with the Cluster Agent.
• The Kubernetes version <= 1.24 (with cgroupV1)

• You have integrated your Cisco AppDynamics SaaS tenant with  Cisco Observability Platform tenant. See Integrate Cisco AppDynamics SaaS and Cisco Cloud Observability.

  

  The account representative helps to integrate the SaaS Controller tenants with  Cisco Observability Platform tenant. 

• You have specified the coexistence.containers.enabled : true setting in the Administration Console. This can be updated by the account representative. 
• You have installed the Cisco AppDynamics Kubernetes collectors and operators. See, Install Kubernetes and App Service Monitoring Using Helm Charts.

Observe Workloads 

You can observe the workloads for a specific application or tier by using the Correlations tab. This tab displays the workloads that are monitored by the cluster collector. This ensures that you get the context of both the high-level (application) data and low-level (workload) data at one place.

Use the Correlations tab to:

• view the available workloads along with their health status for any application or tier.
• traverse to the Cisco Cloud Observability UI from the Cisco AppDynamics SaaS Controller UI to view the workload details.

The Correlations tab is available on the right pane of Application and Tiers & Nodes dashboard. The tab displays all workloads related to the respective application or tier.

To observe the workloads, perform the following:

1. Navigate to the dashboard page
   • Applications > <application-name> > Dashboard > Correlations
   • Tiers & Nodes > <tier-name> > Dashboard > Correlations
2. Click the required workload to open the corresponding workload details page on Cisco Cloud Observability. The UI opens on a new browser tab.
   To view the Workloads list page on Cisco Cloud Observability with the filtered list of workloads corresponding to the specific application or tier, click View workloads in CCO.
3. Troubleshoot the issue based on the Kubernetes entities data on Cisco Cloud Observability.

Use Correlations to Troubleshoot Application Issues  

On an application or tier flowmap, Correlations displays the workloads along with their health. This helps you to understand if a specific workload is causing application issues.

When you click the specific workload, it takes you to the CCO UI, where you get more visibility to the exact issue. The Workloads page provides detailed information that helps you to troubleshoot any issue with the Kubernetes entities. 

The Workloads details page includes:

• the relationship panel to view all the related Kubernetes entities.
• the middle pane that displays all the workload metrics to help you analyse the performance of the workload.
• the right pane to view the properties or attributes of the specific workload.

For more information about the Workloads page, see Workloads.

In the following example of application dashboard, the flowmap shows unhealthy nodes and the Correlations tab displays the orderservice workload as unhealthy. This helps the app owner to find out that there is some issue with a Kubernetes entity. 

To investigate the issue further, the app owner can perform the following from the Correlations tab:

1. Click the orderservice workload.
   The CCO UI is displayed on a new tab.
   
2. Click Health Violations. It shows that there is high memory usage. Also, the Memory Usage vs Limit (%) metric shows that the total memory used is more than the limit.
   This helps in understanding that the issue is with the memory limit.
3. Click Pods in the relationship pane.
   The relationship pane shows the unhealthy pod for the corresponding workload. This helps to update the memory