Service Mesh

Objective

This document provides instructions on how to configure Volterra service mesh and monitor your application services using the service mesh in the VoltConsole. Volterra service mesh provides observability for the application by presenting detailed information on the service interaction for your application. To know more about the service mesh concepts, see Service Mesh.

Using the instructions provided in this document, you can configure service mesh and monitor various details such as metrics, API endpoints, and anomalies.


Prerequisites

  • Volterra Account

    Note: If you do not have an account, see Create a Volterra Account.

  • One or more applications deployed on Volterra site and services configured.

    Note: See K8s Deployment for more information. See Site Management for site creation instructions.

  • In case of a standalone pod such as a traffic generator, set the ves.io/app_type annotation for both the pod and the service for the service mesh to display accurate service graph. For more information on annotations, see Resource Management for vK8s.

Configuration

Configuring Service Mesh and monitoring site deployment from VoltConsole includes the following activities:

Activity Description
Configure Service Mesh Deploy your applications with required configuration for service mesh.
Inspect Service Graph View the site dashboard to check metrics, alerts, reachability, and overall health.
Inspect Service Mesh Dashboard View the site dashboard to check metrics, alerts, reachability, and overall health.
Inspect Service Metrics View the system metrics such as throughput, drop rate, etc.
Inspect Virtual Services View the site status to check the connectivity, software, OS, and scaling status.
Inspect Service Endpoints Check the nodes of your site and their detailed status.
Inspect Service Alerts Check the nodes of your site and their detailed status.
Inspect Service Requests Check the nodes of your site and their detailed status.

Configure Service Mesh

Configuring service mesh requires you to first create a vK8s object and deploy your applications using the kubeconfig associated with the vK8s object.

Step 1: Log into VoltConsole, create vK8s object, and download its kubeconfig.

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

Step 2: Prepare kubernetes manifest file for your application.
  • 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 Volterra to generate 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 Volterra 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.
Step 3: Deploy your application.

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.


Log into VoltConsole and change to your namespace. Select Mesh from the configuration menu and Service Mesh from the options pane. A list of tiles with earch tile representing a service mesh for an application gets displayed with the overall application health, total number of services, and indvidual service health for each listed application.

Click on your application from the displayed list to load the service graph for it.

Inspect Service Graph

The service graph presents an overall graph representation for your application's service interactions. The graph shows service interactions representing services as nodes and interactions as edges. The service request flow is indicated between the edges with arrows

Hover on any node or edge to display a snapshot of the overall application health and metrics.

This example shows the snapshot for an edge.

graph hover
Figure: Service Graph for your Application Services

Click on any node or edge to display a quick information card for that service or interaction.

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 respetively. This example shows the quick view for a node.

graph explore
Figure: Service Graph Explore View for Service

Display service mesh only for one node of your application.

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.

graph explore 2
Figure: Service Graph Explore View for Service

Set time interval, refresh, or filters for your service graph view.
  • Click Last 10 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 10 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 Filter next to the refresh option and select Server or Client or Connection to display filter the display accordingly.
  • Click Table view to display the services in the tabular form.

Inspect Service Mesh Dashboard

The application's dashboard presents an overall monitoring view for your application so that you can inspect overall services, unhealthy services, HTTP errors, latency distribution, and trends for service metrics. Monitor your site dashboard as per the following guidelines:

Open the dashboard view of your service mesh.

Click the Dashboard tab to load the service dashboard.

srv db
Figure: Service Mesh Dashboard

Monitor overall service snapshot and unhealthy services.
  • Check 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 detials on the unhealthy services.
Monitor active alerts from the dashboard view.
  • Check 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.
Monitor the HTTP errors from dashboard.
  • 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 erros for specific HTTP codes using the filter option. The default selected filter is 3xx.
  • Check HTTP Errors as Server to obtain the server erros for HTTP codes 4xx and 5xx.
  • Check HTTP Errors as Client to obtain the client erros for HTTP codes 4xx and 5xx.
Monitor latency, request rate, and throughput for services from dahboard.
  • 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.
Set time interval or a refresh for your dashboard.

Click Last 10 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 10 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.


Inspect Service Metrics

Click the Metrics tab to load the service metrics monitoring view.

The service metrics presents detailed service metrics of request rate, error rate, latency, and throughput. The metrics are displayed in graphs representing the trend over a period of time (default 1 hour). You can select to display client or server trends using the toggle option. You can also select specific nodes or edges to display the metrics for that service or interaction.

Monitor request rate and error rate.
  • 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.

rate metrics
Figure: Rate Metrics

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

Monitor latency trend.
  • 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.

lat metrics
Figure: Latency Metrics

Monitor throughput trends.
  • 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.

updown comb
Figure: Combination of Upstream and Downstream Throughputs

Obtain metrics trend for a specific time interval.

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.


Inspect Virtual Services

Monitor the virtual services for your application.

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

vs
Figure: Virtual Services List


Inspect Service Endpoints

The service endpoints view provides the API endpoints for the services of your application in a graph view with the root and leafs representing the different API paths. You can also obtain the Probability Density Function (PDF) for the Request, Error, Latency, and Throughput (RELT) metrics for the API endpoints.

Load the endpoints monitor view.

Click the Endpoints tab to load the endpoints view.

srv eps
Figure: Service API Endpoints

Switch to PDF view for the endpoint metrics.

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

Hover over any metric to display its PDF value in terms of percenteage, percentile, and mean values.

eps table
Figure: Service API Endpoints PDF view

Note: You can also use the search option to display PDF for a specific service or set of services.

Display probability density for a specific metric of a specific endpoint.

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

eps rr pdf
Figure: Service API Endpoints PDF view

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


Inspect Service Alerts

The service alerts view provides monitoring of the alerts specific to the application services.

Monitor service alerts.

Click the Alerts tab to load the alerts view.

srv alerts
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.


Inspect Service Requests

The service requests view provides monitoring of the requests specific to the application services. This view presents the request trend for your services using sampled HTTP requests over a default or selected time period.

Monitor the requests trend for your services.

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

srv req
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.


Concepts


API References