Package for writing Nagios/Icinga/et cetera plugins in Go (golang).
See https://godoc.org/github.com/Newlode/nagiosplugin.
The general usage pattern looks like this:
func main() {
// Initialize the check - this will return an UNKNOWN result
// until more results are added.
check := nagiosplugin.NewCheck()
// If we exit early or panic() we'll still output a result.
defer check.Finish()
// obtain data here
// Add an 'OK' result - if no 'worse' check results have been
// added, this is the one that will be output.
check.AddResult(nagiosplugin.OK, "everything looks shiny, cap'n")
// Add some perfdata too (label, unit, value, min, max,
// warn, crit). The math.Inf(1) will be parsed as 'no
// maximum'.
check.AddPerfDatum("badness", "kb", 3.14159, 0.0, math.Inf(1), 8000.0, 9000.0)
// Parse a range from the command line and warn on a match.
warnRange, err := nagiosplugin.ParseRange( "1:2" )
if err != nil {
check.AddResult(nagiosplugin.UNKNOWN, "error parsing warning range")
}
if warnRange.Check( 3.14159 ) {
check.AddResult(nagiosplugin.WARNING, "Are we crashing again?")
}
}
In the example above, multiple results were added to the check with AddResult()
. The final state of the check will be determined by its most severe result. By default, result severity is ordered to match the plugin return codes documented in the Nagios plugin developer guidelines. A WARNING result is considered more severe than OK, CRITICAL more severe than WARNING, and UNKNOWN most severe of all. A status policy may be used to modify this default behaviour:
func main() {
check := nagiosplugin.NewCheckWithOptions(nagiosplugin.CheckOptions{
// OK -> UNKNOWN -> WARNING -> CRITICAL
StatusPolicy: nagiosplugin.NewOUWCStatusPolicy(),
})
defer check.Finish()
// Now, any WARNING or CRITICAL results added to the check will be
// considered more severe than any UNKNOWN result.
}
Requires go >= 1.0; tested with versions up to 1.7.