Adjusting the percentage of sampled requests: Agent overhead

To adjust the App Visibility agent performance overhead, modify the number of sampled, or persisted, requests.

The accuracy of metrics is affected by the sampling rate; a lower sampling rate means lower overhead but also lower accuracy.

# The percentage of total requests that will be sampled (i.e. monitored) by the agent. 
# Allowed values: 0-100 inclusive.
# SETTINGS: light is 10, medium is 100, deep is 100
# Default is 100.
sampling.rate=100

Tip

Begin sampling requests at the default value, 100%. Usually, this rate does not noticeably impact application performance, but if traffic is very high and application performance is negatively impacted, lower the sampling rate. In a high-traffic situation, reducing the sample rate does not considerably reduce the accuracy.

You can also tag transaction requests so that transactions with specific headers are always persisted, regardless of the sampling rate.

The agent reads header types from the Agent Policy. For each request type (regardless of sampling), the agent collects the HTTP headers and their values. If the header exists, the agent persists the request, and the App Visibility collector adds the request to the database.

Related topics

Modifying an App Visibility agent policy file to collect and monitor application information

Managing App Visibility agent and confidentiality policies

Persisting tagged transactions

Following up on problematic transactions: escalation

Was this page helpful? Yes No Submitting... Thank you

Comments