This utility displays the extended scheduler tasks information for tree
TREENAME on standard output. This might be useful for scheduler planning
and analysis.
OPTIONS
--config
Reports scheduler's configuration. The values are explained below.
Total collector/monitor leaves: N
Total number of datasources being processed by collector or monitor daemon.
Scheduled leaves by type
Torrus supports arbitrary number of collector types, and this report
shows how many datasources belong to every type. Currently the monitor leaves
are not divided into types.
Least common period: N seconds
The report below shows how the tasks are distributed across the time line,
and the least common period shows the period of this time line.
Tasks execution timeline
This report tells which task and how many datasources are involved in
each task startup event, and how these events are dispersed across the time.
The column Interval shows the time interval from each task execution
event to the next event on the timeline.
--runtime
Reports scheduler's runtime statistics, such as running cycle times,
late starts etc. The meaning of the output values is as follows. Values
that have zero values are usually not printed.
Task: Name, Period: N seconds, Offset: M seconds
Each scheduler task is characterized by its name (usually Collector or
Monitor), period, and timeoffset. Fore example, if period is set to 300
seconds, and offset is 14 seconds, then the task would be executed at
00:00:14, 00:05:14, 00:10:14, and so on for every hour in a day.
N running cycles passed
How many times the task was executed since last reset. The counter
is normally reset after torrus_compilexml(8) successfully
compiles the configuration.
N late starts
How many times the task has started with a delay from its normal schedule.
N too long runs
How many times the task execution time was longer than its period.
N overrun periods
How many periods have been missed because of too long executions.
N missed periods
How many periods were missed because of any reason (such as other tasks
delaying).
Min, Max, Average, Exp Average
The time values are displayed in four columns: Mimimum, Maximum and Average
values since last reset, and Exponential decay value, which may be interpreted
as the average for last several values. With defaults defined in
torrus-config.pl, 95% of the average is calculated from last 3 values.
Running Time
How long the task executes each period.
Late Start
How long the task start was delayed.
Too Long
How much time the too long run took.
RRD Queue
In a multithreaded environment, the RRD files are writen in a background
thread. This value shows the length of the RRD update queue at the beginning
of each update cycle.