Agent Flare

Sending an Agent Flare is not supported for this site.

A flare allows you to send necessary troubleshooting information to the Datadog support team.

This page covers:

A flare gathers all of the Agent’s configuration files and logs into an archive file. It removes sensitive information, including passwords, API keys, Proxy credentials, and SNMP community strings.

The Datadog Agent is completely open source, which allows you to verify the code’s behavior. If needed, the flare can be reviewed prior to sending since the flare prompts a confirmation before uploading it.

Send a flare from the Datadog site

To send a flare from the Datadog site, make sure you’ve enabled Fleet Automation and Remote configuration on the Agent.

To send a remote flare:

  1. From the Fleet Automation page, select an Agent that requires support.
  2. Click Support.
  3. Click Send Support Ticket.
  4. Provide an existing Zendesk support ticket number. If you don’t provide a ticket number, one is created on your behalf.
  5. Enable Debug mode to allow Datadog support staff to troubleshoot your issue faster. The log level is reset to its previous configuration after you send the flare.
  6. If you are troubleshooting your application, enable application tracer level logs to be included in the flare.
  7. Click Send Ticket.
The Send Ticket button launches a form to send a flare for an existing or new support ticket

Send a flare using the flare command

Use the flare subcommand to send a flare. In the commands below, replace <CASE_ID> with your Datadog support case ID if you have one, then enter the email address associated with it.

If you don’t have a case ID, enter your email address used to log in to Datadog to create a new support case.

Confirm the upload of the archive to immediately send it to Datadog support.

PlatformCommand
AIXdatadog-agent flare <CASE_ID>
Dockerdocker exec -it dd-agent agent flare <CASE_ID>
macOSdatadog-agent flare <CASE_ID> or via the web GUI
CentOSsudo datadog-agent flare <CASE_ID>
Debiansudo datadog-agent flare <CASE_ID>
Kuberneteskubectl exec -it <AGENT_POD_NAME> -- agent flare <CASE_ID>
Fedorasudo datadog-agent flare <CASE_ID>
Redhatsudo datadog-agent flare <CASE_ID>
Susesudo datadog-agent flare <CASE_ID>
Sourcesudo datadog-agent flare <CASE_ID>
WindowsConsult the dedicated Windows documentation
HerokuConsult the dedicated Heroku documentation
PCFsudo /var/vcap/jobs/dd-agent/packages/dd-agent/bin/agent/agent flare <CASE_ID>

Dedicated containers

When using Agent v7.19+ and using the Datadog Helm Chart with the latest version or a DaemonSet where the Datadog Agent and Trace Agent are in separate containers, you deploy an Agent Pod containing:

  • One container with the Agent process (Agent + Log Agent)
  • One container with the process-agent process
  • One container with the trace-agent process
  • One container with the system-probe process

To get a flare from each container, run the following commands:

Agent

kubectl exec -it <AGENT_POD_NAME> -c agent -- agent flare <CASE_ID>

Process Agent

kubectl exec -it <AGENT_POD_NAME> -c process-agent -- agent flare <CASE_ID> --local

Trace Agent

kubectl exec -it <AGENT_POD_NAME> -c trace-agent -- agent flare <CASE_ID> --local

Security Agent

kubectl exec -it <AGENT_POD_NAME> -c security-agent -- security-agent flare <CASE_ID>

System probe

The system-probe container cannot send a flare so get container logs instead:

kubectl logs <AGENT_POD_NAME> -c system-probe > system-probe.log

ECS Fargate

When using ECS Fargate platform v1.4.0, ECS tasks and services can be configured to allow access to running Linux containers by enabling Amazon ECS Exec. After enabling Amazon ECS exec, run the following command to send a flare:

aws ecs execute-command --cluster <CLUSTER_NAME> \
    --task <TASK_ID> \
    --container datadog-agent \
    --interactive \
    --command "agent flare <CASE_ID>"

Note: ECS Exec can only be enabled for new tasks. You must recreate existing tasks to use ECS Exec.

PlatformCommand
Kuberneteskubectl exec -n <NAMESPACE> -it <CLUSTER_POD_NAME> -- datadog-cluster-agent flare <CASE_ID>
Cloud Foundry/var/vcap/packages/datadog-cluster-agent/datadog-cluster-agent-cloudfoundry flare -c /var/vcap/jobs/datadog-cluster-agent/config <CASE_ID>

Manual submission

The Agent flare protocol collects configurations and logs into an archive file first located in the local /tmp directory. Manually obtain this file and provide it to support if there are any issues with Agent connectivity.

Kubernetes

To obtain the archive file in Kubernetes, use the kubectl command:

kubectl cp datadog-<pod-name>:tmp/datadog-agent-<date-of-the-flare>.zip flare.zip -c agent

Further Reading

Additional helpful documentation, links, and articles: