Untitled :: Kloudfuse Docs

Kfuse-Knight Agent for Instrumentation-less tracing

The Kfuse-Knight agent collects and aggregates dynamic data on nodes using the Extended Berkeley Packet Filter (eBPF). It runs as a daemonset on a Kubernetes cluster, and can automatically trace protocol-specific network traffic to and from the containers to generate RED metrics. By default, this agent enables tracing for the http protocol. Contact mailto:support@kloudfuse.com to enable collection for other protocols.

The Kfuse Knight pod exposes RED metrics as a Prometheus metrics endpoint; both Prometheus and Datadog can discover and scrape these metrics.

Install the Kfuse-Knight Agent

Create a namespace. kubectl create namespace kfuse-knight

Get the container registry key from kloudfuse <token>.json — uploaded to the page.

Create a k8s secret to use this registry access token. The helm uses it to upgrade and install.

Use either Ubuntu OS or Windows OS.

Ubuntu OS
Windows OS

cat token.json | docker login -u _json_key  --password-stdin https://us-east-1-docker.pkg.dev
kubectl create secret docker-registry kfuse-image-pull-credentials \
        --namespace='kfuse-knight' --docker-server 'us.gcr.io' --docker-username _json_key \
        --docker-email 'container-registry@mvp-demo-301906.iam.gserviceaccount.com' \
        --docker-password=''"$(cat token.json)"''

type token.json | helm registry login -u _json_key  --password-stdin https://us-east-1-docker.pkg.dev
$service_json = (Get-Content token.json).replace("\n", "").replace('"', '\"')
kubectl create secret docker-registry kfuse-image-pull-credentials --namespace=kfuse-knight --docker-server 'us.gcr.io' --docker-username _json_key --docker-email 'container-registry@mvp-demo-301906.iam.gserviceaccount.com' --docker-password=""$service_json""

Get the latest knight chart version.

helm upgrade --install knight oci://us-east-1-docker.pkg.dev/mvp-demo-301906/kfuse-helm/kfuse-knight --version=0.1.0-9688fac -n kfuse-knight

Customizing the Metrics

You can customize the Kfuse-Knight helm values when extracting metrics from logs. This enables you to control the Knight protocols for tracing the necessary information, and to accurately tag it for generated metrics.

The config section of the helm values contains a section of what metrics to extract and how to tag them. You can customize them to fulfill your organization’s requirements. See Extracting and tagging metrics.

The following example specifies what tags to include or exclude when emitting logs. Additionally:

Specify exact field names.
You can specify a Regex pattern.
When specifying both includes and excludes, simply add the field names to each list.
When using JSON formatting, you can extract subfields using the following format:
```
field_name$JSON$<json pointer>
```

Extracting and tagging metrics

config:
  http:
    metrics:
      tags_include:
        - request_type
        - response_code
        - request_headers$JSON$/Accept
        - request_headers$JSON$/Host
      tags_exclude:

yaml

Scraping Metrics

Knight exposes a Prometheus endpoint for its collected metrics.

Recommendation

To automatically enrich the collected metrics with Cloud and Kubernetes labels, use the Datadog agent to scrape the metrics.

In addition to the configurations specified in Collecting Data Using Datadog Agent, add the following configuration in the Datadog agent yaml file.