[dev.icinga.com #5043] Implement range support for warning/critical values in PerfdataValue - Githubissues

icinga-migration commented 11 years ago

This issue has been migrated from Redmine: https://dev.icinga.com/issues/5043

Created by gbeutner on 2013-11-09 09:05:30 +00:00

Assignee: (none) Status: New Target Version: Backlog Last Update: 2016-12-08 00:51:15 +00:00 (in Redmine)

Problem

Performance data ranges are not properly parsed.

The following example from 'check_procs':

procs=1;1:;1:;0;

would mean

warn = "1:" (1 to infinity)
crit = "1:" (1 to infinity)
min = 0
max = null

Specification

Perfdata Value

'label'=value[UOM];[warn];[crit];[min];[max]

Threshold Ranges

A range is defined as a start and end point (inclusive) on a numeric scale (possibly negative or positive infinity).

A threshold is a range with an alert level (either warning or critical). Use the set_thresholds(thresholds *, char *, char *) function to set the thresholds.

The theory is that the plugin will do some sort of check which returns back a numerical value, or metric, which is then compared to the warning and critical thresholds. Use the get_status(double, thresholds *) function to compare the value against the thresholds.

This is the generalised format for ranges:

[@]start:end
Notes:

start ≤ end

start and ":" is not required if start=0

if range is of format "start:" and end is not specified, assume end is infinity

to specify negative infinity, use "~"

alert is raised if metric is outside start and end range (inclusive of endpoints)

if range starts with "@", then alert if inside this range (inclusive of endpoints)

Additional Resources

https://www.monitoring-plugins.org/doc/guidelines.html#THRESHOLDFORMAT https://www.monitoring-plugins.org/doc/guidelines.html#AEN200

Proposal

Find a way to store threshold ranges inside the PerfdataValue class. Evaluate whether these attribute names make sense for inclusive and negative infinity.

warn_upper
warn_lower
crit_upper
crit_lower

Additional thoughts

Graphios doesn't support threshold ranges. https://github.com/shawn-sterling/graphios/blob/master/graphios.py#L351

inGraph parses the ranges https://github.com/NETWAYS/ingraph/blob/master/ingraph/parser.py#L58 and stores upper and lower and additional the type ("inside") - the latter would become cumbersome as time series field in graphite. https://github.com/NETWAYS/ingraph/blob/master/ingraph/parser.py#L132 https://github.com/NETWAYS/ingraph/blob/master/ingraph/parser.py#L224 https://github.com/NETWAYS/ingraph/blob/master/ingraph/model.py#L855

datapoint = Table('datapoint', metadata,
    Column('plot_id', Integer, ForeignKey('plot.id'), nullable=False, primary_key=True, autoincrement=False),
    Column('timeframe_id', Integer, ForeignKey('timeframe.id'), nullable=False, primary_key=True, autoincrement=False),
    Column('timestamp', Integer, nullable=False, primary_key=True, autoincrement=False),
    Column('min', Numeric(precision=20, scale=5, asdecimal=False), nullable=False),
    Column('max', Numeric(precision=20, scale=5, asdecimal=False), nullable=False),
    Column('avg', Numeric(precision=20, scale=5, asdecimal=False), nullable=False),
    Column('lower_limit', Numeric(precision=20, scale=5, asdecimal=False)),
    Column('upper_limit', Numeric(precision=20, scale=5, asdecimal=False)),
    Column('warn_lower', Numeric(precision=20, scale=5, asdecimal=False), nullable=True),
    Column('warn_upper', Numeric(precision=20, scale=5, asdecimal=False), nullable=True),
    Column('warn_type', Enum('inside', 'outside', name='warn_type_enum'), nullable=True),
    Column('crit_lower', Numeric(precision=20, scale=5, asdecimal=False), nullable=True),
    Column('crit_upper', Numeric(precision=20, scale=5, asdecimal=False), nullable=True),
    Column('crit_type', Enum('inside', 'outside', name='crit_type_enum'), nullable=True),
    Column('count', Integer, nullable=False),

This obviously fits into an RDBMS backend, but not Graphite or any other time series database.

Changesets

2015-03-02 12:26:33 +00:00 by jyoung15 af72a60c4b04fb480715c174bb2508b65c38af63

Update PerfdataValue::Parse to ignore invalid warn, crit, min, and max values

... instead of generating exception so that metric values can
still be passed to the graphing backend.
Also update icinga-perfdata tests to reflect these changes.

refs #5043

Signed-off-by: Michael Friedrich <michael.friedrich@netways.de>

2015-03-02 12:52:41 +00:00 by jyoung15 faf6ce1a5f2eb5468e6ebb5db4f7fd08419d2c35

Update PerfdataValue::Parse to ignore invalid warn, crit, min, and max values

... instead of generating exception so that metric values can
still be passed to the graphing backend.
Also update icinga-perfdata tests to reflect these changes.

refs #5043

Signed-off-by: Michael Friedrich <michael.friedrich@netways.de>

Relations:

relates #5043
relates #5043
relates #5043
relates #9461
duplicates #5043
duplicates #5043

icinga-migration commented 11 years ago

Updated by gbeutner on 2013-11-11 07:35:41 +00:00

Target Version set to 0.0.4

icinga-migration commented 11 years ago

Updated by mfriedrich on 2013-11-11 09:28:32 +00:00

Target Version changed from 0.0.4 to 0.0.5

icinga-migration commented 11 years ago

Updated by mfriedrich on 2013-11-12 16:59:28 +00:00

Assigned to set to mfriedrich