Service Mesh

Follow the instructions in the Create vK8s document to create vK8s object and download kubeconfig.

• Prepare a K8s manifest file for your application. Refer to a sample manifest file here.

• Edit the manifest file with the ves.io/app_type labels for your service entry. This is required for Mesh to generate a service mesh for your appl services. The value for this can be the name of the application you are deploying or the namespace name. This label can be set for vk8s service and vk8s container (when the container does not have service associated with it).

The following example applies apiep-doc as the value for ves.io/app_type label for the service part of the manifest.

apiVersion: v1
kind: Service
metadata:
  name: emailservice
  labels:
    app: emailservice
    ves.io/app_type: apiep-doc
  annotations:
    ves.io/proxy-type: HTTP_PROXY
    ves.io/http2-enable: "true"
spec:
  type: ClusterIP
  selector:
    app: emailservice
    ...

Note: The following apply:

• The default value for the ves.io/app_type label is the namespace name.
• In case of a standalone pod such as a traffic generator, set this label for both the pod and the service for the service mesh to display accurate service graph.
• The app_type label can also be applied to the virtual hosts.

• Add Distributed Cloud Services specific annotations to manage resources for vK8s deployments in the deployment manifest. See Resource Management for vK8s for more information. This example adds ves.io/workload-flavor annotation to a deployment manifest:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: emailservice
  annotations:
    ves.io/workload-flavor: medium
spec:
 ...

• Save the manifest.

Use the vK8s kubeconfig and K8s application manifest file obtained in the previous instructions and deploy your application.

kubectl apply -f k8s-app-manifest.yaml --kubeconfig vk8s-kubecfg.yaml

This deploys the application services on the sites selected by the virtual site associated with the vK8s object.

Note: Wait till all the pods are started. For detailed instructions, see the K8s Deployment document.

This example shows the snapshot for an edge.

Figure: Service Graph for your Application Services

Click on the Details and Health tabs to obtain quick information on the current status and health respectively. From within those views, you can click Alerts for nodes or Endpoints for edges to load the alerts or endpoint views respectively. This example shows the quick view for a node.

Figure: Service Graph Explore View for Service

Click the Explore Service option on the quick view to display a filtered view of service mesh only for that service.

You can also click the node or edge to display quick view card for that service or interaction.

Figure: Service Graph Explore View for Service

• Click Last 5 minutes dropdown on the upper right end of the view and select a time interval to inspect your site dashboard for that interval. The default for this is 5 minutes. You can also refresh the status periodically by setting a refresh value using the Refresh every field. Alternatively, click Refresh next to the Last 5 minutes option to manually refresh the status.

Click the Dashboard tab to load the service dashboard.

Figure: Service Mesh Dashboard

• Check the Services part of the dashboard to get a count for the healthy, unhealthy, and total services. You can also click Services in this view to load the service graph view.
• Check Top Unhealthy Services to obtain details on the unhealthy services.

• Check the Active Alerts part of the dashboard to get a count for the critical, major, minor, and total active alerts. You can also click Active Alerts in this view to load the alerts view.
• Check Top Active Alerts by Service to obtain the active alerts per service.
• Check Critical Active Alerts by Service to obtain the critical active alerts per service.

• Check HTTP Errors by Service part of the dashboard to check the HTTP errors per service for both client and server. You can also display errors for specific HTTP codes using the filter option. The default selected filter is 4xx.
• Check HTTP Errors as Server to obtain the server errors for HTTP codes 4xx and 5xx.
• Check HTTP Errors as Client to obtain the client errors for HTTP codes 4xx and 5xx.

• Check Latency Distribution by Service part of the dashboard to check the latency per service.
• Check Request Rate, Latency and Throughput of Services part of the dashboard to check the graph of latency versus request rate. The dot size represents the relative throughput and hover over the dots to see the latency, request rate, and throughput values for the service represented by the dot.

Click Last 5 minutes dropdown on the upper right end of the dashboard and select a time interval to inspect your service dashboard for that interval. The default for this is 5 minutes. You can also refresh the status periodically by setting a refresh value using the Refresh every field. Alternatively, click Refresh next to the Last 10 minutes option to manually refresh the status.

• Click on Request Rate, Error Rate, Error Rate 4XX, and Error Rate 5XX fields on the right side under the Rate section to display the rate trend. Each field has 2 graph bar fields to its left. Select an area of one field and the other area of the second field to display combined graph for in and out drop rate.

Figure: Rate Metrics

Note: The Request Rate trend for the server is displayed by default.

• Click on Latency, Client RTT, Server RTT, and App Latency fields on the right side under the Latency section to display the latency trend. Each field has 2 graph bar fields to its left. Select an area of one field and the other area of the second field to display combined graph for those metrics.

Figure: Latency Metrics

• Click on Upstream Throughput or Downstream Throughput fields on the right side under the Throughput section to display the throughput trend. Each field has 2 graph bar fields to its left. Select an area of one field and the other area of the second field to display combined graph for them.

Figure: Combination of Upstream and Downstream Throughputs

Click Last 1 hour dropdown on the upper right end of the dashboard and select a time interval to inspect your site dashboard for that interval. The default for this is 1 hour and maximum allowed interval is 24 hours. You can customize the interval by selecting the Custom option and choosing date range. This can also be set graphically by adjusting the controls beneath the main graph.

Click the Virtual Services tab to load the virtual services view.

Figure: Virtual Services List

Click the API Endpoints tab to load the endpoints view.

Figure: Service API Endpoints

Click the Table option to load the PDF for the RELT metrics for all of the endpoints. The following metrics are displayed:

• Request size and response size
• Request rate and error rate
• Latency with and without data
• Response throughput

The following operations are available in the table:

• Hover over any metric to display its PDF value in terms of percentage, percentile, and mean values.
• Click Download Swagger to see get the swagger file for the endpoint in JSON format.
• Use the Search pull-down menu to display PDF for a specific service or set of services.
• Click the gear icon, ⚙, to choose which columns to view.

Figure: Service API Endpoints PDF view

Click on the PDF of any metric in the Table view to load its probability density trend in a graph.

Figure: Service API Endpoints PDF view

Note: You can also switch to other metrics using the X Axis drop-down to display its graph.

Click the Alerts tab to load the alerts view.

Figure: Service Alerts

The active alerts are displayed by default. Use the toggle selection to load all alerts. You can also set a time interval in the active alerts view to display alerts over a specific period of time. Click > for any alert entry to display details in JSON format.

Click the Requests tab to load the view for the trend of sampled HTTP requests.

Figure: Service Requests

The requests are displayed in a graphical trend as well as in a list for the default or specific time interval. Click > for any listed request to display detailed information in JSON format.

You can also apply filters to display the trend for specific HTTP codes. For example, de-select all and select only 2xx to display the requests for HTTP code 2XX.