Remstats Home


  About remstats
  architecture
  Release Notes
  FAQ
  Conventions
  Prerequisites
  Installation
  Configuration
      access
      alerts
      alert-destination-map
      alert-template-map
      alert-templates
      archives
      availability
      colors
      customgraphs
      dbi-connects
      dbi-selects
      environment
      general
      groups
      hosts
      host templates
      html
      links
      ntops
      oids
      pages
      page-templates
      remotepings
      run
      run-stages
      rrds
      scripts
      times
      tools
      views
      view templates
  Configuration Tools
  Servers
  Collectors
  Updater
  Monitors
  pagemakers
  run-remstats2
  CGIs
  troubleshooting
  do-traceroutes
  Miscellany
  Thank-you
  Index


  Live Data

  Remstats was written by
  Thomas Erskine at the
  CRC in Canada and now
  at SourceWorks.

[Top] [Prev] [Next]

Configuration - Alerts

The alerts config-file is used by the alert-monitor to decide who to send alerts for problems. The rrds and hosts config-files decide when an alert needs to be raised, and these lines tell who gets the alert.

Each line is in seven parts, most of which are patterns, e.g.:

	warn	* MISC UPTIME 0 0 uptime-alerts
	error   silverlock.dgim.crc.ca  * * 600 900 test-alerts
	error   news.crc.ca * * 600 900 news-alerts
	critical * * * 600 900 critical-alerts

The first "word" is the status, as decided by the alert-monitor. The second word is a regex to match against the hostname. The third is a regex to match against the rrdname. The fourth is a regex for the variablename. The fifth is the minimum time for the alert condition to be present before an alert can be triggered. The sixth is the interval after sending an alert before another will be sent. The seventh is an alert-destination as specified in the alert-destination-map.

Note: The seventh used to be an alert program and there was an eighth which was an address, of a form appropriate to the alert program. This has been rolled into the alert-destination-map to make it more flexible.

If the current condition matches the status, host, rrd, and variable, then alert-monitor has to look at the times. If this is a new condition (i.e. it was in OK status previously), then an alert won't be triggered until after the minimum time has passed. This avoids transient problems being reported. If you want these to be reported, then set it to zero. If this is an old alert, then an alert won't be triggered until the interval time has passed since the previous alert. If the interval is 0 (zero) then there will only be one alert at the start-time.


[Top] [Remstats] [SourceWorks] [RRDtool] [SourceForge]
Last updated Fri Jun 22 13:37:42 GMT 2007 by <terskine@users.sourceforge.net>.