This the multi-page printable view of this section. Click here to print.

Return to the regular view of this page.

  • 1:
    • 1.1:

    Serverless Agents


    The serverless environment: As cloud platforms have evolved, both the convenience and the abstraction levels have increased simultaneously and new agent models are required.

    For example, with Amazon’s ECS and EKS, users remain in charge of managing the underlying virtual host machines. In environments like Fargate, however, the hosts are implicitly allocated by the cloud provider and users simply run their containers without allocating, configuring, or having any knowledge of the underlying compute infrastructure.

    While this “container as a service” model is convenient, it can introduce risk, as many users leave the containers unattended and don’t monitor for security events inside them that can exfiltrate secrets, compromise business data, and increase their AWS/cloud provider costs. In addition, it is not possible to install a standard agent in an environment where you do not have access to a host.

    For these reasons, Sysdig has introduced a new “serverless agent” that can be deployed in such container-based cloud environments.

    Available Platforms

    1 -

    AWS Fargate Serverless Agents

    Check the Overview for an explanation of when and why to use serverless agents in “container-as-a-service” cloud environments.


    The Sysdig serverless agent provides runtime detection through policy enforcement with Falco. At this time, the serverless agent is available for AWS Fargate on ECS. It is comprised of an orchestrator agent and (potentially multiple) workload agents.

    • The Sysdig serverless orchestrator agent is a collection point installed on each VPC to collect data from the serverless workload agent(s) and to forward them to the Sysdig backend. It also syncs the Falco runtime policies and rules to the workload agent(s) from the Sysdig backend.

    • The Sysdig serverless workload agent is installed in each task and requires network access to communicate with the orchestrator agent.

    Installation: For Fargate ECS

    For Fargate ECS, the two components of the serverless agent are installed separately.

    • For the orchestrator agent, Sysdig provides a yaml to use as a CloudFormation Template which you can deploy through the AWS Console. You need one orchestrator deployment per VPC in your environment which your organization wants to secure.

    • For the workload agents, you need one workload agent per Fargate task definition. (If you have ten services and ten task definitions, each needs to be instrumented.)

      We assume your services use an existing CFT and you will install the workload agent using an automated process which will instrument all the task definitions in your CFT.


    On the AWS side:

    • AWS CLI configured and permissions to create and use an S3 bucket.

    • Permissions to upload images to repos, deploy CloudFormation Templates (CFTs), and create task definitions for Fargate

    • The Fargate tasks you want to instrument with the Sysdig serverless agent

    • Two subnets that can connect with the internet. (Your service on Fargate must reach the orchestrator agent, and the orchestrator agent must reach the internet to communicate with Sysdig’s back end.)

    • A NAT gateway or an AWS Internet Gateway so the orchestrator agent can reach to Sysdig.

    On the Sysdig side:

    Install the Orchestrator Agent

    1. Obtain the Sysdig Orchestrator Agent yaml to be used as the CloudFormation Template source.

      For more information on CloudFormation (CFN), see AWS documentation.

    2. Deploy the orchestrator agent for each desired VPC, using CloudFormation.

      The steps below are an outline of the important Sysdig-related parts.

      1. Log in to the AWS Console. Select CloudFormation and Create Stack with new resources and specify the orchestrator-agent.yaml as the Template source.

      2. Specify the stack details to deploy the orchestrator agent on the same VPC where your service is running.

        Stack name: self-defined

        Sysdig Settings

        • Sysdig Access Key: Use the agent key for your Sysdig platform.

        • Sysdig Collector Host: (default); region-dependent in Sysdig SaaS; custom in Sysdig on-prem.

        • Sysdig Collector Port: 6443 (default), or could be custom for on-prem installations.

        Network Settings

        • VPC Id: Choose your VPC.

        • VPC Gateway: Choose the type of Gateway in your VPC to balance the load of the orchestrator service.

        • Subnet A & B: These depend on the VPC you choose; select from the drop-down menu.

        Advanced Settings

        • Sysdig Agent Tags: Enter a comma-separated list of tags (eg. role:webserver,location:europe) Note: tags will also be created automatically from your infrastructure’s metadata, including AWS, Docker, etc.

        • Sysdig Orchestrator Agent Image:


        • Check Collector SSL Certificate: Default: true. False means no validation will be done on the SSL certificate received from the collector, used for dev purposes only.

        • Sysdig Orchestrator Agent Port: 6667 (default). Port where the orchestrator service will be listening for instrumented task connections.

      3. Click Next, complete the stack creation, and wait for the deployment to complete (usually less than 10 minutes.)

      4. In Output, take note of the OrchestratorHost and OrchestratorPort values.

        As of Serverless Agent Version 2.2.0, the orchestrator port can be configured via either SYSDIG_ORCHESTRATOR_PORT (default 6667, which is hardcoded into other parts of the CFT) or the SysdigOrchestratorAgentPort (alternate) parameter in the CloudFormation template.

    If AWS Internet Gateway is used (as opposed to a NAT Gateway), uncomment the line AssignPublicIp: ENABLED in the orchestrator.yaml.

    Install the Workload Agents

    Option 1: Automated Process (Kilt)

    1. Prerequisite: Have the orchestrator agent deployed in the appropriate VPC and have the Orchestrator Host and Port information handy.

    2. Download the appropriate installer for your OS.

      These set up Kilt, an open-source library mechanism for injection into Fargate containers.

    3. Create a macro for the serverless worker agents, using the installer. Any service tagged with this macro will have the serverless worker agent(s) added and Fargate data will be collected.

      1. Log in to AWS CLI.

      2. Create a CFN macro that applies instrumentation. You will need the outputs from previous task. Example: 

        ./installer-linux-amd64 cfn-macro install -r us-east-1 MySysdigMacro $OrchestratorHost $OrchestratorPort
    4. Add the macro you created to the CFT that you use for your own service at the root.

      Use: Transform: MySysdigMacro.

      All new deployments of that template will be instrumented.

      Defining Entrypoint and Command in CFT:

      In this step, the macro will go over the original CloudFormation Template looking for ContainerDefinitions to instrument. This includes replacing the original entry point for the Sysdig entry point, so ideally the CFT should explicitly describe Entrypoint and Command. Otherwise, the macro will try to pull the image to retrieve a default Entrypoint and Command, which only works if the image is publicly available. Otherwise the pull will fail and the instrumentation will not be completed.

    5. Complete!

      When instrumentation is complete, Fargate events should be visible in the Sysdig Secure Events feed.

    Option 2: Automated Process Terraform

    Alternatively, you can use the Sysdig Provider on the Terraform registry to deploy the workload agents, as described below.

    1. Prerequisite: Have the orchestrator agent deployed in the appropriate VPC and have the Orchestrator Host and Port information handy.

    2. Set up the Sysdig Terraform provider:

      terraform {
        required_providers {
          sysdig = {
            source = "sysdiglabs/sysdig"
            version = ">= 0.4.0"
      provider "sysdig" {
        sysdig_secure_api_token = var.sysdig_access_key
    3. Pass the container-definition, orchestrator host, and orchestrator port to the sysdig_fargate_workload_agent data source.


      • The input container definitions must be in JSON format, following CloudFormation naming conventions.

      • When deploying the orchestrator agent with CFT, the port is always 6440.

      • When deploying the orchestrator agent with CFT, the host is the collector endpoint for your Sysdig region; e.g. "" for US East. For other regions, see SaaS Regions and IP Ranges.

        data "sysdig_fargate_workload_agent" "instrumented" {
          container_definitions = file("container-definitions/hello.json")
          sysdig_access_key = var.sysdig_access_key
          workload_agent_image = ""
          orchestrator_host = module.sysdig_orchestrator_agent.orchestrator_host
          orchestrator_port = module.sysdig_orchestrator_agent.orchestrator_port
    4. Include the instrumented JSON in your Fargate task definition:

      resource "aws_ecs_task_definition" "fargate_task" {
        network_mode = "awsvpc"
        requires_compatibilities = ["FARGATE"]
        container_definitions = "${data.sysdig_fargate_workload_agent.instrumented.output_container_definitions}"
    5. Complete!

      When instrumentation is complete, Fargate events should be visible in the Sysdig Secure Events feed.


    Environment Variables for Workload Agents

    The most commonly used variables include:

    • SYSDIG_ORCHESTRATOR: (required) use the output OrchestratorHost value
    • SYSDIG_ORCHESTRATOR_PORT: (required) use the output from OrchestratorPort
    • SYSDIG_EXTRA_ARGS: use to pass any extra arguments needed to the instrumentation
    • SYSDIG_EXTRA_CONF: use to add any extra configuration needed to the instrumentation
    • SYSDIG_LOGGING: add this variable and set to debug if needed. Any other value will print minimal logs at startup. See also: Manage Serverless Agent Logs

    Manual Instrumentation for Workload Agents

    In some cases, you may prefer not to use the Kilt or Terraform installers and instead to either:

    • Instrument a Task Definition, or
    • Instrument a Container Image (rare)

    Option A: Instrument a Task Definition

    1. Review the installation overview and prerequisites.

    2. Install the Orchestrator Agent, as described above. Take note of the OrchestratorHost and OrchestratorPort values.

    3. Instrument the CloudFormation task definition to deploy the workload agent manually.

    4. Add a new container to your existing CloudFormation task definition.

      • For the example, we will use SysdigInstrumentation for the image name.
      • The entrypoint/command fields can be left empty
      • Obtain the image from
    5. Edit the containers you want to instrument.

      • Mount volumes from SysdigInstrumentation

      • Add SYS_PTRACE capability to your container. See AWS Documentation for detail if needed.

      • Set /opt/draios/bin/instrument as the entry point of your container.

        Note: if you are using the entry point in your image, then prepend it: /opt/draios/bin/instrument,your,original,entry,point

    6. Configure the Sysdig instrumentation through environment variables.

    Example Task Definitions

    CloudFormation Task Before Instrumentation `

      Type: AWS::ECS::TaskDefinition
        NetworkMode: awsvpc
          - FARGATE
        Cpu: 2048
        Memory: 8GB
        ExecutionRoleArn: !Ref TestAppExecutionRole
        TaskRoleArn: !Ref TestAppTaskRole
          - Name: App
            Image: !Ref TestAppImage
            - /bin/ping
          - Key: application
            Value: TestApp

    Task After Manual Instrumentation

      Type: AWS::ECS::TaskDefinition
        NetworkMode: awsvpc
          - FARGATE
        Cpu: 2048
        Memory: 8GB
        ExecutionRoleArn: !Ref TestAppExecutionRole
        TaskRoleArn: !Ref TestAppTaskRole
          - Name: App
            Image: !Ref TestAppImage
    ### Added for manual instrumentation
            - /opt/draios/bin/instrument
    ### End added propertis
            - /bin/ping
    ### Added for manual instrumentation
                - SYS_PTRACE
            - SourceContainer: SysdigInstrumentation
              ReadOnly: true
            - Name: SYSDIG_ORCHESTRATOR # Required
              Value: <host orchestrator output, region dependent>
            - Name: SYSDIG_ORCHESTRATOR_PORT # Required
              Value: "6667"
            - Name: SYSDIG_ACCESS_KEY
              Value: ""
            - Name: SYSDIG_COLLECTOR
              Value: ""
            - Name: SYSDIG_COLLECTOR_PORT
              Value: ""
            - Name: SYSDIG_LOGGING
              Value: ""
          - Name: SysdigInstrumentation
            Image: !Ref WorkloadAgentImage
    ### End of added properties
          - Key: application
            Value: TestApp

    Option B: Instrument in a Container Image

    Alternatively, you can include the workload agent in your container at build time. To do so, update your Dockerfile to copy the required files:

    ARG sysdig_agent_version=latest
    FROM$sysdig_agent_version AS workload-agent
    FROM my-original-base
    COPY --from=workload-agent /opt/draios /opt/draios

    Prepend the /opt/draios/bin/instrument command to the entrypoint of your container:

    ENTRYPOINT ["/opt/draios/bin/instrument", "my", "original", "entry", "point"]

    Note about Logs

    To separate instrumentation logs from your workload logs, the instrument entrypoint will attempt to send logs to localhost:32000, where they will be mixed with the workload logs.

    • If you want to separate instrumentation logs in a manual install, we advise using Option A.
    • To avoid log spam, you can set SYSDIG_LOGGING to silent.

    See also: Manage Serverless Agent Logs

    Upgrade: For Fargate ECS

    To upgrade the serverless agents, you install a second version of both components then kill all the running tasks and restart with the new version. You then delete and clean up the old CloudFormation and Kilt residuals.
    1. Obtain the Sysdig Orchestrator Agent yaml to be used as the CloudFormation Template source.

      At this time, the yaml metadata does not specify the agent version, but this link always downloads the latest file.

    2. Perform the steps to Install the Orchestrator Agent.

      Note that in step 2.4, the OrchestratorHost and OrchestratorPort values will be unique.

    3. Perform the steps to Install the Workload Agents (if using Kilt. Workload agents installed with Terraform are automatically updated to _latest.)Manual instrumentation requires redoing the manual process.)

      In step 4, assign a unique name to your macro (Transform: MyV2SysdigMacro).

      You now have two versions of the serverless agent components. When you are ready to switch from the earlier version, proceed with the next step.

    4. Stop all running tasks and use CloudFormataion to delete the earlier stack. Redeploy the new stack with the updated CFT. (Transform: MyV2SysdigMacro)

    5. Clean up macros: Delete the previous kilt macro using the installer:

      ./installer-linux-amd64 cfn-macro delete MySysdigMacro

    1.1 -

    Manage Serverless Agent Logs

    As of serverless agent version 2.2.0, task logs and instrumentation logs are handled separately.

    • Task logs continue to be with whatever log setup you have on your task container.

    • Instrumentation logs go to a separate log group created by the macro installer:


      At this time, the log group name cannot be edited. By default, the logging level is set to info.

    Logging Environment Variables

    Set Logging Level

    • SYSDIG_LOGGING is used for instrumentation log level. Default = info. The full list of options are: none | critical| error | warning |notice | info | debug | trace

    Log Forwarding

    • SYSDIG_ENABLE_LOG_FORWARD Default = true
    • SYSDIG_LOG_LISTEN_PORT Default = 32000
    • SYSDIG_LOG_FORWARD_ADDR Default = localhost:32000. Where the logwriter of the workload agent listens. Can be set to any other TCP endpoint.

    Advanced Configurations

    • SYSDIG_BUFFER_SIZE: Receiving buffer size in bytes for the TCP listener, which can be increased. Default 1024.
    • SYSDIG_DEADLINE_SECONDS: Connection deadline, which can be increased if clients are taking longer to connect. Default 3.

    See other serverless agent environment variables here.