What is a Probe Group?
Multiple services run on one machine - It is not uncommon to have more than one service running on a single machine, and thus to have multiple devices in Int representing that machine. One might be for SNMP traffic, another for HTTP, a third for FTP. All probe the same IP address; they just look at different aspects of the host. Int allows you to collect any number of devices probing the same IP address into a single ?probe group?.
One IP address, multiple probes - A probe group represents the multiple services in a single device. In the map view, a probe group uses a single icon; opening the Info window for the icon allows you to see the individual services. In the list views, a probe group is a single list item with a "+" (Windows/Linux) or triangle (Mac). Clicking the + or triangle expands the view to reveal the individual services.
Member probes - Individual probes within a probe group are called ?member probes?.
A probe group with the Info Window opened to reveal member probes.
The same probe group in list view, expanded to reveal member probes.
A probe group counts as a single device against your license, no matter how many member probes it contains.
Polling and State
Poll interval and probe state - Each member probe is polled at its own poll intervals and maintains its own state.
Shows the worst state of the group - The state of the probe group is the ?worst? of the states of the member probes. The state of the probe group will be evaluated and changed, if necessary, each time a member probe changes state.
How Do You Create a Probe Group?
There are several ways to create a probe group.
Group selected devices - If you are creating one from existing devices, it?s quite easy. Simply select the items you want to include in the group, right-click on one of the items, and choose ?Group?. The individual devices will disappear from the map, and probe group will appear in place of one of them. The probe group will have its own icon and label, based on the defaults set in the Device Defaults panel of the Server Settings.
Create an empty group - It is also possible to create one from scratch. Begin by adding a device in the usual way; when it is time to select a probe type, find the Probe Group section in the probe picker and select the Probe Group probe, then continue in the usual way. This will create an empty probe group. You can then use the method in the previous paragraph to group existing devices into the group or use the Info window to add new member probes directly to the group. (See below.)
Probe Group Info and Status
A probe group has an Info Window and a Status Window, just like any other device. Each type of window has additional functionality for probe groups.
Edit the group from the Info Window - In the Info Window for probe groups, there is an extra panel with a list of the member probes, if any. Name, probe type, condition, and state for each member probe are included. Use the ?+? and ?-? at the bottom of the window to add new member probes or delete existing member probes. (See above for an example of the info window.)
View status of the group - The status window for probe groups displays not only the status of the probe group, but a list of member probes with the state, probe type, and first line of the label of the member probe. Expand a list item to see complete status information for the member probe.
A probe group status window, with the section for one member probe expanded.
Manipulating Member Probes
The device list view also allows you to get and set the properties and characteristics of individual member probes. Right-click on a member probe in an expanded view of the probe group to get the customary contextual menu for a device.
The contextual menu for a member probe; a few functions normally
available for devices (such as the interfaces window or Set
Address) are not available for member probes, but most are.
Attach a notifier to group, to member probes, or both - You may attach notifiers to probe groups in the same way as you attach them to any other device. These notifiers will be triggered based on the state of the probe group. In addition, any or all member probes may have their own notifiers triggered by their individual states. The notifiers attached to the probe group and the notifiers attached to the member probes are independent of each other. Each will be triggered as appropriate without regard to the other. This means that it Is conceivable that you could receive two notifiers for the same event; one because the state of a member probe changed and one because that state change caused the state of the probe group to change.
Event log shows notifications by probe name or probe broup name - When notifications from a member probe are logged to the Event Log, they will include ?(in ProbeGroupName)? to indicate that ?ProbeGroupName? is the name of the probe group in which the probe is a member. Similarly, when notifications from a probe group are logged to the Event Log, they will include ?(due to MemberProbeName)? to indicate that the probe group changed status as a result of a change to ?MemberProbeName?.
Include Probe Group name, Probe Name, or both in Notifier message - When editing the message for notifiers, you can choose to include ?Probe Group? and/or ?Member Probe? in the message to indicate the same information.
The Configure Notifier panel?s Edit Message dialog;
note the Probe Group and Member Probe options.
The Control Probe
Click the star to make a member probe into a control probe - It is possible to set a member probe as a ?control probe?. To do so, in the Info Window for the probe group, click the star next to the member probe you wish to set as the control probe.
Control probe determines interface visibility, notifier behavior - There are two implications of setting a member probe as control probe. First, the interfaces for that member probe will become visible as the interfaces for the probe group. Second, if the control probe goes DOWN, no subsequent notifiers for other member probes will be sent until the control probe is UP again.
First SNMP probe encountered automatically becomes control probe - When you group existing devices into a probe group, the first SNMP probe encountered during the grouping will automatically be set as the control probe. No control probe will be set automatically when adding probes directly to a probe group.
The info window from our first example, this time with a control probe selected. Note the star. Note, too, that the probe group now shows the SNMP Traffic probes interfaces.
Ungrouping a Probe Group
You may find that you wish to ungroup a probe group into its component devices. To do so, select the probe group, right-click, and choose ?Ungroup?. Any devices that were collected into the probe group as a result of the ?Group? operation will return to their original locations, with their original icons and labels. Any devices resulting from probes added directly to the probe group will be given a new location, icon, and label based on the defaults set in the Device Defaults panel of the Server Settings.
One fact that sometimes surprises our users is that probe groups currently do not inherit any label information from member probes. Thus, setting the label to use something like ?SNMP SysName? will not have the expected effect, because the probe group is not an SNMP device. It is likely that in a future version of Int, we will change this behavior such that the probe group label can inherit information from the control probe.
What is Shared Polling?
Under certain situations, Int will allow two or more devices to ?share? polling; that is, to send only one query and to share the result among the devices doing the sharing, thus reducing network traffic. Int also counts shared devices as a single device against the license limit.
Probes to which Shared Polling Applies
Int will automatically set up shared polling for ICMP/Ping or SNMP devices which are polling the same IP address and which are identical in every way. (Probe, poll interval, community string, etc.)
How Shared Polling Takes Effect
Int will check to see if devices can share polling after certain events such as loading a map, adding a device, changing a probe type, etc. When it sets devices to share polling, all but one device are subtracted from the number of devices counted against the license?s device limit. Similarly, when it finds that two or more devices no longer share polling, they will be re-added to the number of devices counted against the license?s device limit.
You can tell a device is sharing polling by looking at the last line of the status window of the device. If there is a number in parentheses (e.g. ?(2)?), that number represents the number of devices sharing polling with that device, including the device itself.
The last line of this status window indicates that there are four devices,
Including this one, sharing SNMP Traffic polling to nitro.dartware.com.
Pasting device or loading a map can temporarily affect your device count - the way in which Int goes about looking for the opportunity to share polling means that in some cases, you may need more devices to load a map or paste devices than you will need once the operations have completed. While this happens only infrequently, you will need to be aware of it if you are near your device limit. Future versions of Int are expected to be more forgiving so that this is not necessary.
"Shared Polling"-like Situations
There are several other probes which may ?share? polling, but which are not treated as shared polling and which do not reduce device counts in the same way.
Command-line and NT Services probes which start to poll while an identical poll is already underway will simply register to receive the results of the poll already underway rather than continue with an additional poll.
TCP probes, when they start to poll, will check to see if an identical poll is already underway and will simply wait to receive the results of the poll already underway rather than continue with an additional poll.