The Nobl9 Agent
Nobl9 provides the ability to run an Agent to retrieve SLI metrics from configured data sources and send the data back to the Nobl9 backend for processing and error budget calculation. You can run the Agent in your Kubernetes cluster or as a Docker container.
The Agent is a lightweight application that executes the queries defined for your Nobl9 SLOs. Queries are written in the language supported by the data source in question and executed via native APIs. The interval at which the queries are executed varies by data source, but in most cases is one minute (refer to the Sources section of the documentation for more details about the resolution of the queries).
Why Use the Nobl9 Agent?
Nobl9 users can choose between a Direct or Agent configuration when connecting to a data source.
A Direct configuration requires users to enter their authentication credentials (API key, token, etc.). These values are encrypted and safely stored in Nobl9.
With an Agent configuration, the user passes their credentials when launching the Agent, and those credentials are not stored in the Nobl9 backend. What's more, you have access to the Agent's logs, which makes troubleshooting easier (see the Agent Troubleshooting page). For this reason, we recommend making your initial connection to a data source with an Agent configuration.
You can also use the Nobl9 Agent to collect and return data if your company's firewall blocks outbound connections. You still need to open a port for the Nobl9 Agent to send data back to Nobl9, but it removes the need for the Nobl9 application to make direct calls to your environment.
Getting Started with the Nobl9 Agent
You can deploy the Nobl9 Agent in any Kubernetes cluster or Docker environment.
A Docker environment on your local machine with proper firewall access is enough for testing purposes. However, we do not advise using it past the initial test, as the data flow will stop when your machine sleeps.
Make sure that you are using the latest version of the Agent, so you have access to all the most recently introduced features. For details on the Agent releases, refer to the Agent Release Notes.
Creating the Agent
Creating the Agent in the UI
Follow these steps to create your Agent in the UI:
- Go to Integrations.
- Click the button.
- Select the relevant source and choose Agent.
For more Integration-specific details, refer to the Agent Configuration in the UI section in each Integration's subpage.
Creating the Agent through
To create your Agent through
sloctl, you can apply the YAML file to deploy the Agent by using the
sloctl apply command, for example:
sloctl apply -f ./agent.yaml
For more information, refer to the
sloctl User Guide. For the Source-specific YAML definition of Agents, check the Integrations section - you can find them in the Agent Using CLI - YAML section in each Integration subpage.
You can create only one Agent in one YAML file.
If you describe infrastructure as code, you might also consider creating Agents via Terraform. You can find more details in our Terraform documentation.
Deploying the Nobl9 Agent
Deploying the Agent in Kubernetes
When you add a new Agent via UI or
sloctl, Nobl9 automatically generates a Kubernetes configuration in YAML and a Docker command line for you:
If you configure your Agent in the UI, these configurations are generated for you immediately.
If you create your Agent via YAML, you must go to the Integrations > Sources tab and find the respective Agent configuration tabs under Details.
If you have a running Kubernetes cluster, you can copy and paste the generated YAML into your Kubernetes configuration to deploy the Nobl9 Agent in your cluster.
Be sure to swap in your own credentials (e.g.,
<CREDENTIALS_FILE>); instructions in the UI specify what credentials need to be passed, and these differ depending on the data source.
Deploying the Agent in a Docker Container
When you add a new data source through the Nobl9 UI, a Docker command line is generated automatically for you to use to deploy the Agent. The following is a generic example of what you can see in the UI:
Be sure to swap in your own credentials (e.g.,
<CREDENTIALS_FILE>); instructions in the UI specify what credentials need to be passed.
Checking the Agent Connection in the UI
To verify that the Agent has successfully connected to Nobl9, check for a valid timestamp in the Last Connection field in the UI.
Note that the Connected status does not indicate that the Agent is connected to the specified data source, only that it has successfully established a connection to the Nobl9 backend.
If you're using Prometheus, you can also check the connection status by scraping in from Agent's
/health endpoint. Check Agent Metrics for more details.
Updating the Nobl9 Agent
If you need to update your Nobl9 Agent, you must apply the new config file with the latest image of the Agent in your Kubernetes cluster or Docker container.
Go to Integrations, find the relevant source on the list, and click the update button:
The new Docker and Kubernetes configuration will be applied for you immediately, copy-paste the generated file and apply it to your Kubernetes cluster or run it in your Docker container:
You can go to Docker hub to check the latest Docker images of the Nobl9 Agent.
Exposing the Agent to Metrics
To expose the Agent's internal metrics, you must make the following changes in your Kubernetes cluster or Docker invocation:
To deploy the Agent, add the environment variable
N9_METRICS_PORT, specifying the port on which you would like to expose the metrics (e.g.,
N9_METRICS_PORT=9876). These metrics are then available for scraping at
If you deploy the Agent behind a firewall, make any changes required to allow your metrics system to scrape that port.
The Agent must be able to open connections to
https://app.nobl9.com/api/inputto send data to Nobl9.
Customizing Query Delay
Nobl9 Agent supports a
QUERY_DELAY environment variable that allows to set a customized delay for queries when pulling the data from the data source.
Nobl9, by default, tries to pull data from the previous minute. Sometimes the data in the observability platform may not be available, and, as such, specifying the query delay variable allows you to pull data from a time further than 1 minute in the past.
The following table specifies the list of the
QUERY_DELAY environment variable names specific to each data source:
|Data source||Variable name|
|Google Cloud Monitoring|
The variable value must be specified in the Golang Duration format. For more information, refer to the
The following is an example of YAML configuration with the
- name: agent-container
- name: N9_CLIENT_ID
- name: N9_CLIENT_SECRET
- name: DYNATRACE_TOKEN
- name: DYNATRACE_QUERY_DELAY
Agent Timestamps Cache Persistence | Nobl9 Documentation
Agent Troubleshooting | Nobl9 Documentation
Agent Metrics | Nobl9 Documentation