nagios.responsetime 1.0.2

A Nagios plugin that collects response times from logs.

A Nagios plugin that collects response times from logs.

Usage: check_responsetime --logfile /path/to/logfile

The plugin remembers the timestamp of the latest log entry it has seen on each run, and on the next run only considers log entries that are more recent.

Log format

The supported log format is as follows:

<timestamp> "<http-verb> <path> <http-version>" <responsetime>

• timestamp: %d/%b/%Y:%H:%M:%S %z

• http-verb: GET, POST

• path: the requested URL path

• http-version: e.g. HTTP/1.1

• responsetime in seconds, e. g. 0.5

Example:

17/May/2011:14:11:18 +0200 "GET /index.html HTTP/1.1" 0.289

An example configuration for the nginx server looks like this:

http {
    log_format timing '$time_local "$request" $upstream_response_time';

    server {
        access_log /path/to/timing.log timing;
    }
}

Future directions

One obvious additional feature is the ability to define thresholds (e.g. “a mean responsetime above 3 seconds is WARN, above 5 CRITICAL”)

Contents

• Log format

• Future directions

  • Changes

    • 1.0.2 (2011-11-14)

    • 1.0.1 (2011-09-04)

    • 1.0 (2011-07-01)

    • 1.0rc1 (2011-06-10)

Changes

1.0.2 (2011-11-14)

• Fixed divide by zero bug when there is only a single entry (#9332).

1.0.1 (2011-09-04)

• Fixed bug when there are no log entries for the current interval (#9332).

1.0 (2011-07-01)

• Fixed bug with incomplete log file lines (“invalid literal for float(): -“).

1.0rc1 (2011-06-10)

• First release.