Chart: Sysdig

Sysdig is a unified platform for container and microservices monitoring, troubleshooting, security and forensics. Sysdig platform has been built on top of Sysdig tool and Sysdig Inspect open-source technologies.

Table of Contents

• Introduction
• Prerequisites
• Installing the Chart
• Uninstalling the Chart
• Configuration
• Resource profiles
• Node Analyzer
• GKE Autopilot
• On-Premise backend deployment settings
• Using private Docker image registry
• Modifying Sysdig agent configuration
• Upgrading Sysdig agent configuration
• How to upgrade to the latest version
• Adding custom AppChecks
• Support

Introduction

This chart adds the Sysdig agent for Sysdig Monitor and Sysdig Secure to all nodes in your cluster via a DaemonSet.

Prerequisites

• Kubernetes 1.9+ with Beta APIs enabled

Installing the Chart

First of all you need to add the Sysdig Helm Charts repository:

$ helm repo add sysdig https://charts.sysdig.com/

To install the chart with the release name sysdig-agent, run:

$ helm install --namespace sysdig-agent sysdig-agent --set sysdig.accessKey=YOUR-KEY-HERE --set sysdig.settings.collector=COLLECTOR_URL sysdig/sysdig --set nodeAnalyzer.apiEndpoint=API_ENDPOINT

To find the values:

• YOUR-KEY-HERE: This is the agent access key. You can retrieve this from Settings > Agent Installation in the Sysdig UI.
• COLLECTOR_URL: This value is region-dependent in SaaS and is auto-completed on the Get Started page in the UI. (It is a custom value in on-prem installations.)
• API_ENDPOINT: This is the base URL (region-dependent) for Sysdig Secure and is auto-completed on the Get Started page. E.g. secure.sysdig.com, us2.app.sysdig.com, eu1.app.sysdig.com.

After a few seconds, you should see hosts and containers appearing in Sysdig Monitor and Sysdig Secure.

Tip: List all releases using helm list

Uninstalling the Chart

To uninstall/delete the sysdig-agent deployment:

$ helm delete --namespace sysdig-agent sysdig-agent

Tip: Use helm delete --namespace sysdig-agent --purge sysdig-agent to completely remove the release from Helm internal storage

The command removes all the Kubernetes components associated with the chart and deletes the release.

Configuration

The following table lists the configurable parameters of the Sysdig chart and their default values.

Node Image Analyzer parameters (deprecated by nodeAnalyzer)

Specify each parameter using the --set key=value[,key=value] argument to helm install. For example,

$ helm install --namespace sysdig-agent sysdig-agent \
    --set sysdig.accessKey=YOUR-KEY-HERE,sysdig.settings.tags="role:webserver\,location:europe" \
    sysdig/sysdig

Alternatively, a YAML file that specifies the values for the parameters can be provided while installing the chart. For example,

$ helm install --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

Tip: You can use the default values.yaml

Resource profiles

For ease of use, some predefined resource profiles are available:

• small

requests:
  cpu: 1000m
  memory: 1024Mi
limits:
  cpu: 1000m
  memory: 1024Mi

• medium

requests:
  cpu: 3000m
  memory: 3072Mi
limits:
  cpu: 3000m
  memory: 3072Mi

• large

requests:
  cpu: 5000m
  memory: 6144Mi
limits:
  cpu: 5000m
  memory: 6144Mi

• custom

By setting “custom” or any value other than the ones defined above, you can create your own custom profile to match your requirements by setting the appropriate values in resources object.

See Tuning Sysdig Agent for more info.

Node Analyzer

The Node Analyzer is deployed by default unless you set the value nodeAnalyzer.deploy to false.

The Node Analyzer daemonset contains three containers, each providing a specific functionality. This daemonset replaces the (deprecated) Node Image Analyzer daemonset.

See the Node Analyzer installation documentation for details about installation, and Running Node Analyzer Behind a Proxy for proxy settings.

Node Image Analyzer

See the Image Analyzer Configmap Options for details about the available options, and the Node Image Analyzer documentation for details about the Node Image Analyzer feature.

The node image analyzer (NIA) provides the capability to scan images as soon as they start running on hosts where the analyzer is installed. It is typically installed alongside the Sysdig agent container.

On container start-up, the analyzer scans all pre-existing running images present in the node. Additionally, it will scan any new image that enters a running state in the node. It will scan each image once, then forward the results to the Sysdig Secure scanning backend. Image metadata and the full scan report is then available in the Sysdig Secure UI.

Host Analyzer

See the Host Scanning Configuration Options for details about installation options, and the Host Scanning documentation for details about the Host Scanning feature.

The host analyzer provides the capability to scan packages installed on the host operating system to identify potential vulnerabilities. It is typically installed as part of the Node Analyzer which in turn is installed alongside the Sysdig Agent.

The host analyzer works by inspecting the files on the host root filesystem looking for installed packages and sending them to the Sysdig backend. It performs this operation by default once a day and its schedule can be configured as described below. Likewise, the list of directories to be examined during each scan can be configured.

Benchmark Runner

See the Benchmarks documentation for details on the Benchmark feature. The Benchmark Runner provides the capability to run CIS inspired benchmarks against your infrastructure. Benchmark tasks are configured in the UI, and the runner automatically runs these benchmarks on the configured scope and schedule. Note: if nodeAnalyzer.benchmarkRunner.includeSensitivePermissions is set to false, the service account will not have the full set of permissions needed to execute oc commands, which most checks in CIS Benchmark for OS4 require.

GKE Autopilot

Autopilot is an operation mode for creating and managing clusters in GKE. With Autopilot, Google configures and manages the underlying node infrastructure for you.

To deploy the Sysdig agent in GKE clusters running in Autopilot mode, run:

$ helm install --namespace sysdig-agent sysdig-agent --set sysdig.accessKey=YOUR-KEY-HERE --set sysdig.settings.collector=COLLECTOR_URL sysdig/sysdig --set gke.autopilot=true

When the flag gke.autopilot=true gets true, the chart configuration is overridden as follows:

• nodeAnalyzer.deploy=false
• ebpf.enabled=true
• ebpf.settings.mountEtcVolume=false
• daemonset.annotations='autopilot\.gke\.io/no-connect="true"'
• daemonset.affinity=null'

So, on GKE Autopilot clusters:

• The NodeAnalyzed is not deployed,
• The ebpf is enabled and the etcVolume is not mounted,
• The daemonset affinity is set to null,
• The daemonset annotation is set to enable the Agent to run on autopilot (required from GKE).

On-Premise backend deployment settings

Sysdig platform backend can be also deployed On-Premise in your own infrastructure.

Installing the agent using the Helm chart is also possible in this scenario, and you can enable it with the following parameters:

For example:

$ helm install --namespace sysdig-agent sysdig-agent \
    --set sysdig.accessKey=YOUR-KEY-HERE \
    --set collectorSettings.collectorHost=42.32.196.18 \
    --set collectorSettings.collectorPort=6443 \
    --set collectorSettings.sslVerifyCertificate=false \
    sysdig/sysdig

Using private Docker image registry

If you pull the Sysdig agent Docker image from a private registry that requires authentication, some additional configuration is required.

First, create a secret that stores the registry credentials:

$ kubectl create secret docker-registry SECRET_NAME \
  --docker-server=SERVER \
  --docker-username=USERNAME \
  --docker-password=TOKEN \
  --docker-email=EMAIL

Then, point to this secret in the values YAML file:

sysdig:
  accessKey: YOUR-KEY-HERE
image:
  registry: myrepo.mydomain.tld
  repository: sysdig-agent
  tag: latest-tag
  pullSecrets:
    - name: SECRET_NAME

Finally, set the accessKey value and you are ready to deploy the Sysdig agent using the Helm chart:

$ helm install --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

You can read more details about this in Kubernetes Documentation.

Modifying Sysdig agent configuration

The Sysdig agent uses a file called dragent.yaml to store the configuration.

Using the Helm chart, the default configuration settings can be updated using sysdig.settings either via --set sysdig.settings.key = value or in the values YAML file. For example, to eanble Prometheus metrics scraping, you need this in your values.yaml file::

sysdig:
  accessKey: YOUR-KEY-HERE
  settings:
    prometheus:
      enabled: true
      histograms: true

$ helm install --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

Upgrading Sysdig agent configuration

If you need to upgrade the agent configuration file, first modify the YAML file (in this case we are increasing the metrics limit scraping Prometheus metrics):

sysdig:
  accessKey: YOUR-KEY-HERE
  settings:
    prometheus:
      enabled: true
      histograms: true
      max_metrics: 2000
      max_metrics_per_process: 400

And then, upgrade Helm chart with:

$ helm upgrade --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

How to upgrade to the last version

First of all ensure you have the lastest chart version

$ helm repo update

In case you deployed the chart with a values.yaml file, you just need to modify (or add if it’s missing) the image.tag field and execute:

$ helm upgrade --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

If you deployed the chart setting the values as CLI parameters, like for example:

$ helm install \
    --namespace sysdig-agent \
    sysdig-agent \
    --set sysdig.accessKey=xxxx \
    --set ebpf.enabled=true \
    sysdig/sysdig

You will need to execute:

$ helm upgrade \
    --namespace sysdig-agent \
    sysdig-agent \
    --set sysdig.accessKey=xxxx \
    --set ebpf.enabled=true \
    --set image.tag=<last_version> \
    sysdig/sysdig

Adding custom AppChecks

Application checks are integrations that allow the Sysdig agent to collect metrics exposed by specific services. Sysdig has several built-in AppChecks, but sometimes you might need to create your own.

Your own AppChecks can deployed with the Helm chart embedding them in the values YAML file:

customAppChecks:
  sample.py: |-
    from checks import AgentCheck

    class MyCustomCheck(AgentCheck):
        def check(self, instance):
            self.gauge("testhelm", 1)

sysdig:
  accessKey: YOUR-KEY-HERE
  settings:
    app_checks:
      - name: sample
        interval: 10
        pattern: # pattern to match the application
          comm: myprocess
        conf:
          mykey: myvalue

The first section dumps the AppCheck in a Kubernetes configmap and makes it available within the Sysdig agent container. The second one configures it on the dragent.yaml file.

Once the values YAML file is ready, we will deploy the Chart like before:

$ helm install --namespace sysdig-agent sysdig-agent -f values.yaml sysdig/sysdig

Automating the generation of custom-app-checks.yaml file

Sometimes editing and maintaining YAML files can be a bit cumbersome and error-prone, so we have created a script for automating this process and make your life easier.

Imagine that you have custom AppChecks for a number of services like Redis, MongoDB and Traefik.

You have already a values.yaml with just your configuration:

sysdig:
  accessKey: YOUR-KEY-HERE
  settings:
    app_checks:
      - name: myredis
        [ ... ]
      - name: mymongo
        [ ... ]
      - name: mytraefik
        [ ... ]

You can generate an additional values YAML file with the custom AppChecks:

$ git clone https://github.com/sysdiglabs/charts.git
$ cd charts/sysdig
$ ./scripts/appchecks2helm appChecks/solr.py appChecks/traefik.py appChecks/nats.py > custom-app-checks.yaml

And deploy the Chart with both of them:

$ helm install --namespace sysdig-agent sysdig-agent -f custom-app-checks.yaml -f values.yaml sysdig/sysdig

Adding prometheus.yaml to configure promscrape

Promscrape is the component used to collect Prometheus metrics from the sysdig agent. It is based on Prometheus and accepts the same configuration format.

This file can contain relabelling rules and filters to remove certain metrics or add some configurations to the collection. An example of this file could be:

global:
  scrape_interval: 15s
  evaluation_interval: 15s
scrape_configs:
  - job_name: 'prometheus' # config for federation
    honor_labels: true
    metrics_path: '/federate'
    metric_relabel_configs:
      - regex: 'kubernetes_pod_name'
        action: labeldrop
    params:
      'match[]':
        - '{sysdig="true"}'
    sysdig_sd_configs:
      - tags:
          namespace: monitoring
          deployment: prometheus-server

sysdig_sd_configs allows to select the targets obtained by Sysdig agents to apply the rules in the job. Check how to configure filtering in sysdig documentation .

Adding additional volumes

To add a new volume to the sysdig agent.

In order to pass new config maps or secrets used for authentication (for example for Prometheus endpoints) you can mount additional secrets, configmaps or volumes. An example of this could be:

extraVolumes:
  volumes:
    - name: sysdig-new-cm
      configMap:
        name: my-cm
        optional: true
    - name: sysdig-new-secret
        secret:
        secretName: my-secret
  mounts:
    - mountPath: /opt/draios/cm
      name: sysdig-new-cm
    - mountPath: /opt/draios/secret
      name: sysdig-new-secret

Support

For getting support from the Sysdig team, you should refer to the official Sysdig Support page.

In addition to this, you can browse the documentation for the different components of the Sysdig Platform:

• Sysdig Monitor
• Sysdig Secure
• Platform Documentation
• Monitor Documentation
• Secure Documentation