Stackstate-NTP Check Integration

Overview

The Network Time Protocol (NTP) integration is enabled by default and reports the time offset from an ntp server every 15 minutes. When the local agent’s time is more than 15 seconds off from the StackState service and the other hosts that you are monitoring, you may experience:

  • Incorrect alert triggers
  • Metric delays
  • Gaps in graphs of metrics

Setup

Installation

No installation steps are required for this integration.

Configuration

  1. The ntp check is enabled by default. If you would like to make any changes to the configuration, move ntp.yaml.default to ntp.yaml and edit:

    init_config:
    
    instances:
      - offset_threshold: 60
    
  2. Restart the agent

Configuration Options

  • host (Optional) - Host name of alternate ntp server, for example pool.ntp.org
  • port (Optional) - What port to use
  • version (Optional) - ntp version
  • timeout (Optional) - Response timeout

Validation

To validate your installation and configuration, restart the agent and execute the info command. The output should contain a section similar to the following:

Checks
======
  [...]
  ntp
  -----
      - instance #0 [OK]
      - Collected 1 metric & 0 events

Data Collected

Metrics

ntp.offset
(gauge)
The time difference between the local clock and the NTP reference clock.
shown as second

Events

No events are included with this integration.