Monitoring Netagent

Aggregate and summarize networking metrics for the Access Tier

This article describes features that are only available in the Banyan Enterprise edition.
This article describes features that require Banyan Netagent v1.37+.

Overview

The Netagent binary, which is the core component of the Banyan Access Tier, is instrumented to generate metrics to be collected by a StatsD daemon; the daemon will then generate aggregates and relay them to a monitoring backend.

Enabling metrics collection

To enable metrics collection, set the statsd and stats_address parameters in the Netagent config.yaml file:

Parameter Default Value Description
statsd false Turn on metrics collection using statsd
statsd_address 127.0.0.1:8125 Address to send statsd messages: “hostname:port” for UDP, “unix:///path/to/socket” for UDS

Once configured, the following metrics will be sent to the statsd daemon:

Metric Name Metric Type Description
banyan.transmit_rate COUNT Bytes transmitted
banyan.receive_rate COUNT Bytes received
banyan.connections COUNT Number of TCP connections
banyan.errors COUNT Number of failures to connect to the upstream server
banyan.unauthorized_attempts COUNT Number of attempted accesses blocked by policy
banyan.response_time HISTOGRAM Time to respond to a HTTP request
banyan.decision_time HISTOGRAM Time to process a TCP request

Netagent aggregates TX/RX byte counts and sends them to statsd daemon roughly every 13 seconds. For all metrics except TX/RX byte counts, Netagent sends statsd messages to the daemon immediately, where they are received and aggregated.

Every metric is tagged with descriptors that allow you to disaggregate telemetry data as needed:

Tag Name Tag Value
site_name Name of Access Tier that generated the metrics
host_name Linux host running Netagent, representing a specific instance of the Access Tier
service_id ID of the Banyan Service for the metric

Metric tags and the HISTOGRAM metric type are supported explicitly by Datadog, but may not work with other monitoring tools. The Response Time metric is calculated only for Hosted Website services.

Datadog Integration using DogStatsD

DogStatsD is a metrics aggregation service bundled with the Datadog Agent. When you run the Datadog agent alongside Netagent on the Access Tier machines, you can use DogStatsD to collect network metrics for visualization in your Datadog dashboards.

Leveraging Tags for per Service or per Access Tier metrics

Metrics sent to Datadog contain tags that allow filtering at per Service or per Access Tier level. In order to filter metrics at an Access Tier level, navigate to the Metrics Explorer and filter on site_name={Access Tier}.

In order to filter metrics at a Banyan service level, navigate to the Metrics Explorer and filter on service_id={Service Name}.



Last modified: Aug 05, 2021