This post is also available in: French
Do you know the meta-services in Centreon?
Often misunderstood, the meta-services are rarely used. But they are a real added-value for those who know how to manage them. The information that meta-services produce can bring a good knowledge of your IT. This information concern real-time monitoring and performance graphs. This article, based on our expertise and experience, aims to present the meta-services and explain how to use and configure them.
A meta-service is a service based on performance data of several services to reach a status (ok, warning, critical) and news performance data. The idea is to aggregate the performance data of several services to reach relevant information:
- either generate an alert
- or generate a new graph
As a reminder, performance data are collected by Centreon that transforms them into graph. These are, for example, the CPU occupancy rate, the free space on the partition, the HTTP response time…
In order to better understand the meta-services utility, these are practical case-studies.
They are as representative as possible to the use of meta-services.
- Total number of people connected:
when your application is split over several servers or several firewalls, you can measure the number of connected users on each application/server/firewall/network equipment.
- bandwidth used:
You have several Internet service providers with different service level and different bandwidth. You can measure the bandwidth used for each routers. The meta-services allow you to sum all bandwidth used and visualize the evolution of all bandwidth used over the time.
- bandwidth used:
You are a Internet service provider. You offer services based on your own box. You monitor the bandwidth used individually by each customer. These customer can be external (companies) or internal (other entity of your company). Through meta-services, you can get the average bandwidth consumed and whether the average changes over time. This allows you to change your offer based on the needs of your customers.
- Power consumption of servers:servers can measure their own power consumption and provide this information through MIBs SNMP. If you can measure them, you can sum all power consumptions.
As a reminder, meta-services can aggregate data only is they relate to the same thing (same measurement unit). As an example, you can not combine the indicators on the RAM and the CPU.
To define meta-services, go to Configuration –> Services –> Meta-service –> Add.
Then, you have two methods to define them :
- by matching SQL: all services which have the same name are interpreted in the meta-service
- by unit selection of each services
These two methods will be detailed later in this article. Whichever one you selected, the following parameters are common meta-service:
- Meta Service Name
- Output format string (printf-style)
This output have to be compatible with the command “printf” that allow you to display the result. Most of the time, the value is written as following :
“‘information description : %f”. So, %f is replaced by value calculated.
Example : TOTAL FW CONNEXION: %f
- Warning Level
Above this value, the meta-service will be in Warning status.
- Critical Level:
Above this value, the meta-service will be in Critical status.
- Calculation Type
A calulation type is used on data performance of services and the result of this type is calculated. The calculation type available are :
- Average: the average of all values
- Sum: the sum of all values
- Min: minimum of all values
- Max: maximum of all values
A meta-service is also a… service!
It has the same parameters:
- Check Period: monitoring period
- Max Check Attempts: max number of attempts before confirmation of warning
- Normal Check Interval: interval between 2 tests
- Retry Check Interval: interval between 2 tests when confirmation of warning
- Notification Enabled: activation/deactivation of notifications
- Linked Contact Groups: groups of contacts to notify
- Notification Interval: interval between 2 notifications
- Notification Period: notification period
- Notification Type: status for which we notify some contacts
- Graph Template: applied graph model
- Status: meta-service active or non
- Comments: free comment
Meta-service by SQL Matching
Adding a meta-service by SQL Matching allows you to apply the formula chosen for a metric to all services whom the same has the extension SQL.
As an example:
- Selection Mode: SQL Matching
- SQL LIKE-clause expression: nbr-connect
- Metric: connexion
By entering the above parameters, all services named “nbr-connect” and having a metric (or performance data) “connection” will be included in the calculation.
You can replace some characters by the character “%” and thus, you can manage following case:
- “all services beginning by cpu” ==> “cpu%”
- “all services finishing by fw_connexion_number” ==> “%fw_connexion_number”
Meta-service by unit choice of services
This choice allows you to select exactly what services are to be included in the calculation of meta-services. This is the most common. The configuration is done in two steps:
- Definition of meta-service like above but getting some elements:
- Selection Mode: service list
- SQL LIKE-clause expression: to leave blank
- Metric: to leave blank
- To save meta-service
- On the far right of the table, click on the icon for the new created meta-service
- Click on add
- Choose the host, then the service and the metric
- Repeat the operation as often as necessary
Visualisation of meta-services
You can view the meta-services on the page Monitoring–>Services–>Meta-services
Meta-services can express additional information to standard indicators. They are the result of a formula applied to performance data services. We recommend that you use graphics available to improve control of your IT.
Incoming search terms:
- centreon meta service
- meta services