Prometheus exporter and Grafana template

NV_Exporter Setup:

To run the exporter as Python program

• Clone the repository
• Make sure you installed Python 3 and python3-pip:

$ sudo apt-get install python3
$ sudo apt-get install python3-pip

• Install the Prometheus Python client:

$ sudo pip3 install -U setuptools
$ sudo pip3 install -U pip
$ sudo pip3 install prometheus_client requests

To run the exporter and prometheus as a container

It's easier to start NeuVector exporter as a container. The following section describe how to start the exporter in the Docker environment. A kubernetes sample yaml file, nv_exporter.yml, is also included.

Modify both docker-compose.yml and nv_exporter.yml. Specify NeuVector controller's RESTful API endpoint CTRL_API_SERVICE, login username CTRL_USERNAME, password CTRL_PASSWORD, and the port that the export listens on through environment variables EXPORTER_PORT. Optionally, you can also specify EXPORTER_METRICS to a comma-separated list of metric groups to collect and export. It's highly recommanded to create a read-only user account for the exporter.

Metric groups:

• summary - overall NeuVector status
• conversation - total bytes for every conversation between workloads
• enforcer - enforcer CPU and memory usage
• host - host memory usage
• admission - number of allowed and denied Kubernetes admission requests
• image_vulnerability - number of high and medium vulnerabilities for every scanned registry image
• container_vulnerability - number of high and medium vulnerabilities for every service, reporting a single pod's status per service (excluding service mesh sidecars)
• log - data for the latest threat, incident, and violation logs (latest 5 logs each)

Environment Variables

Variable	Description	Default
CTRL_API_SERVICE	NeuVector controller REST API service endpoint	nil
CTRL_USERNAME	Username to login to controller REST API service	admin
CTRL_PASSWORD	Password to login to controller REST API service	admin
EXPORTER_PORT	The port that the export is listening on	nil
ENFORCER_STATS	For the performance reason, by default the exporter does NOT pull CPU/memory usage from enforcers. Enable this if you want to see the metrix in the dashboard	0

In native docker environment

Start NeuVector exporter container.

$ docker-compose up -d

• Open browser, go to: [exporter_host:exporter_port] (example: localbost:8068)
• If you can load the metric page, the exporter is working fine.

Add and modify the exporter target in your prometheus.yml file under scrape_configs:

scrape_configs:
  - job_name: prometheus
    scrape_interval: 10s
    static_configs:
      - targets: ["localhost:9090"]
  - job_name: nv-exporter
    scrape_interval: 30s
    static_configs:
      - targets: ["neuvector-svc-prometheus-exporter.neuvector:8068"]

Start Prometheus container.

$ docker run -itd -p 9090:9090 -v $(pwd)/prometheus.yml:/etc/prometheus/prometheus.yml --name prometheus prom/prometheus

• After deployed Prometheus, open browser and go to: [prometheus_host:9090] (example: localhost:9090)
• On the top bar go to Status -> Targets to check exporter status. If the name is blue and State is UP, the exporter is running and Prometheus is successfully connected to the exporter.
• On the top bar go to Graph and in the Expression box type nv to view all the metrics the exporter has.

In Kubernetes

Start NeuVector exporter pod and service.

$ kubectl create -f nv_exporter.yml

Create configMap for Prometheus scrape_configs.

$ kubectl create cm prometheus-cm --from-file prom-config.yml

Start Prometheus pod and service.

$ kubectl create -f prometheus.yml

Grafana Setup:

• Start Grafana container. "docker run" example,

$ sudo docker run -d -p 3000:3000 --name grafana grafana/grafana

• After deployed Grafana, open browser and go to: [grafana_host:3000] (example: localhost:3000)
• Login and add Prometheus data source from Configurations -> Data Sources
• find the + on the left bar, select Import. Upload NeuVector dashboard templet JSON file.

Metrics

Metrics	Comment
nv_summary_services	Number of services
nv_summary_policy	Number of network policies
nv_summary_pods	Number of pods
nv_summary_runningWorkloads	Number of running containers
nv_summary_totalWorkloads	Total number of containers
nv_summary_hosts	Number of hosts
nv_summary_controllers	Number of controllers
nv_summary_enforcers	Number of enforcers
nv_summary_disconnectedEnforcers	Number of disconnected enforcers
nv_summary_cvedbTime	Vulnerability database build time
nv_summary_cvedbVersion	Vulnerability database version
nv_host_memory	Memory usage of nodes (by node id)
nv_controller_cpu	CPU usage of controllers (by controller id)
nv_controller_memory	Memory usage of controllers (by controller id)
nv_enforcer_cpu	CPU usage of enforcers (by enforcer id)
nv_enforcer_memory	Memory usage of enforcers (by enforcer id)
nv_conversation_bytes	Network bandwidth of applications
nv_admission_allowed	Number of allowed admission control requests
nv_admission_denied	Number of denied admission control requests
nv_image_vulnerabilityHigh	Number of vulnerabilities of high severity (by image id)
nv_image_vulnerabilityMedium	Number of vulnerabilities of medium severity (by image id)
nv_container_vulnerabilityHigh	Number of vulnerabilities of high severity (by service name)
nv_container_vulnerabilityMedium	Number of vulnerabilities of medium severity (by service name)
nv_log_events	Lists of security events
nv_fed_master	Shows the status of all the connected clusters to the federated master
nv_fed_worker	Shows the status of the cluster to the federated master