Many different statistics can be collected and monitored by OverTime.
Types of collections can be categorized by the type of data to be collected.
The following categories will be explained below: I/O, User Specified and
PingTime.
To collect interface traffic statistics for Input/Output type devices
like routers, switches and hubs, simply "otadd" the device specifying
just the community string and the devices address. OverTime then queries the
device looking for I/O interfaces that may be operational to add to the list of
collections. The interface must be administratively Up, be reporting a non-zero
interface speed that is also less than 400Gb and actually be able to be queried
for the SNMP OIDs ifInOctets and ifOutOctets. If the interface provides the
SNMP values for ifInErrors and ifOutErrors, then these are also collected. If
the device supports CiscoAvgBusy5 MIB OID this is also included in the list of
collections.
OverTime creates the collection entries ready for the SNMP collector.
If there were no previous collections and RRD file for the
interface/instance, then the default sample_time value is used for the
frequency of collection and the creation of the RRD. (If sample_time is not
specified in the file overtime.cfg a default of 300 seconds or 5 minutes is
used. See Below.)
OverTime creates all necessary directories for the new device. A directory is
created in the web server’s area (under OverTime) to hold the graphic files
created.
OverTime creates a directory (under OverTime\data) to hold the RRD
files (1 RRD per interface or instance).
OverTime creates a master configuration file for the device. This plain
text file contains the list of interfaces being monitored, one per line. For
I/O collections, each line has four values on it with the first three blanks
separating the 4 values and takes the form:
Interface-Number Code Threshold “Title
Interface-Number is the SNMP Instance Number for this interface;
Code is one of a 0, 1 or 2;
Threshold is a user modifiable threshold that draws a line on the
resulting graphs if the values plotted approach the threshold;
“Title is a user modifiable title for the graph;
The threshold and title are initially set from values taken from the
device.
The default threshold value is the interface speed for interfaces OR
100% for CPU.
The default titles are the interface names.
OverTime updates/creates the master html page (which by default is
index.html) in the web servers OverTime directory. The devices are listed on
this page in alphabetical order (case sensitive). The name chosen for the
device is either 1) its SNMP sysName or, if that is blank, the IP address or
the IP address specified.
When otadd is run, you can also specify non-IO* collections for this
device that are lists of SNMP Object IDs (OIDs) and instance numbers, then
these collections are setup in a similar manner to I/O collections except:
In the master configuration file, a group of lines is added for each
non-IO collection. I.E. if a collection for an OID for Disk Space Used is
requested, then a group of lines would be added where the first line indicated
that Disk Space Used OID was being collected. Each subsequent line in the group
would relate to an instance of this (probably a disk drive) and the group would
be terminated by a line with nothing on it. When the results are viewed through
your browser, this group would appear on the device’s main page as a single
graph, with a line for each instance (disk drive).
When otadd is run, you can collect both I/O and non-IO by specifying
"" "" for the first OID and instance pair to collect all IO
as well as the other OID and instance pair that are specified.
* I/O Collections are collections on ifInOctets, ifOutOctets,
ifInErrors, ifOutErrors, sysUptime and CiscoAvgBusy5.
* non-I/O collections other than those listed as I/O.
When you run otadd, you can
also specify non-SNMP collections be created. These collections use a script,
command or executable that you provide to collect any type of statistic and
have it added to the OverTime environment. An example of this for collecting
web server response times is shown later. With this type
of collection you can monitor non-SNMP values or anything that your OverTime
server can collect statistics on!
In conjunction with the additional product, PingTime, OverTime can
display ping response times from a Cisco router to any IP address. (See
PingTime later). PingTime collections are performed by the PingTime
application.