Monitoring
With the NetEye.Cloud Monitoring service you can monitor the availability of your hosts and services using check commands.
Monitored Objects¶
There are two main types of objects that monitoring works with:
Hosts: Physical or virtual devices that correspond to a “computer”. They have a CPU, memory and storage (which can be measured), and run processes launched by the user or the system itself. Hosts can be “Up”, “Down” or Unreachable, depending on whether they’re functioning or not. Examples include:
Computers (servers, laptops and tablets)
Virtual machines
Network devices (switches, routers)
IT infrastructure (security alarms, smart locks)
IoT devices (temperature sensors, water detector)
Host objects provide a mechanism to group services that are running on the same physical device.
Services: The software programs and processes that run on hosts, like Active Directory, or network services (HTTP/S, SNMP, SSH). They run faster or slower depending on the current performance of their host (as long as their host is “Up”), and even crash and need to be restarted.
The service states are:
OK: The service is working correctly
WARNING: The service is working, but its performance has degraded past a certain, definable point
CRITICAL: The service’s performance is highly degraded, for instance it responds slowly to user input
UNKNOWN: The service’s state can’t be determined (for example, it doesn’t respond)
Modern host operating systems will typically include one or more services that are dedicated to monitoring, such as reporting CPU load, memory or storage capacity used, and network throughput. These are called check commands, and they return the service state and other result values. User-provided check commands called plugins can also be installed.
We can check the status and performance of a service at any given moment by looking at the check results using:
Views, which show the most recent return values
Dashboards, which show multiple views as panels in a larger screen
Interactive Dashboards, which continually store those check results and let you compare them with previously recorded values
Host and Service Templates¶
Because there can be many different types of hosts that need different check commands and other parameters, it’s more efficient to implement an inheritance system of templates than repeatedly re-enter the same information for each host.
Templates allows you to quickly define a new host by choosing the most relevant template and then only making the few necessary changes to differentiate that host from others, such as host name and IP address.
The same logic also applies to services (service templates) and commands (command templates). You can find more details about all three templates in the On-Premise version of the NetEye Guide.
Within NetEye.Cloud there are already many pre-defined templates to use, so it’s more important to learn what they are and how they are used rather than how to create new ones.
How to Add a Host or Service¶
If you need to add a specific host, command or service, that information can also be found in the On-Premise NetEye guide:
How to See Your Hosts and Services¶
In NetEye.Cloud you can view all the hosts in your monitored environment by following in the left side panel. Each host is displayed with a brief summary including:
Its name
Its status
How long the current state has been that way
Additional check or plugin output beyond the state
If a host is down it will be highlighted in red so it is quickly noticed.
When you have a large number of hosts they cannot all be displayed on one page, NetEye.Cloud provides page controls that let you determine how many hosts to show on one page, how to navigate from one page to another, sort the hosts in the list, and quickly search for hosts by name.
You can see the list of your services just like the host list by following .
Check Commands and Plugins¶
A plugin is the monitoring code that exists on the monitored object that actually computes the result and returns the desired value. The check command runs when directed on the NetEye.Cloud side and calls the plugin via a network connection to request the check and to collect the returned value(s). A check command and plugin working together are typically just called a check.
Although you can write your own commands, most of the time you will just select one that already exists in NetEye.Cloud, either internally derived from Icinga’s library, or from a third party library, called external commands. Both types typically have a large number of arguments you can change (or leave blank) directly in NetEye.Cloud.
Fig. 188 Commands on the left, External Commands on the right¶
Each command has a name, and the Arguments tab (or Preview tab for external commands) shows the relationship between the command line arguments and NetEye.Cloud/Icinga variables you can use to populate them.
To select and use an external command, you can search online to find out how to install the associated plugin on the host device and use the arguments to obtain the check values you’re interested in.
Reachability and Dependencies¶
A reachable host is one that can “answer” if you talk to it, and an unreachable host doesn’t answer. This is determined by using a command called hostalive (ping) that sends a network request to the host. If the request is answered, the host is reachable at that moment.
But a host may not answer either because (1) it’s down or (2) it’s up but the network between the host and the monitor is not working. It’s not always easy to tell the difference. For instance, if other hosts on the same subnet are reachable, then the host is likely down. If they aren’t, it’s more likely a network problem.
When a host is down, any services running on that host will also not be working. And any other hosts that need to access services on the first host will have problems as well. These dependencies can be stated explicitly so that when a “parent” host is marked down in the monitoring interface, any “child” hosts and services that depend on it can be immediately marked as Unreachable.
Fig. 189 A host marked as Unreachable¶
Services can also be marked as unreachable. In fact, services can serve as both parents and children of either other hosts or other services.
Dependencies can especially help root cause analysis of a problem, because if a large number of hosts or services stop working, you can concentrate first on the “Down” objects before the “Unreachable” ones.
It can also mitigate the chaos when there are many dependent hosts: Child objects can be set to suppress notifications and further checks, so that a single problem doesn’t result in being inundated by tens of thousands of messages.
