This post is also available in: French
For this 5th edition, the event was much bigger than usual. The meeting was held on Saturday, November 22nd in the heart of Bastille (Paris), in a loft that wears the same name.
The day started with the presentation of the development progress of Centreon 3.
Presentation of Centreon 3.
Broadly speaking, the R&D team of Centreon is renewing the user interface completely : configuration management, organizations, environments, custom views, detailed host views, data structure… It is hard to explain everything thoroughly, but here are some key topics :
The main technical changes of Centreon :
- The Centreon interface is totally modular. All the Centreon views are now modules, it is therefore possible to update modules separately at different release paces, no more hard version dependencies between Centreon and Centreon Broker or Centreon Engine !
- creation of new daemon that is centralized and modular, as replacement for centcore.
- For this module, we use ZeroMQ and LibSSH. this daemon will be in charge of configuration deployment, ACL computing, autodiscovery and whatnots…
- Compatibility break with Nagios and NDOutils; though it is possible to implement new data sources (elastiksearch, influxDB …)
- New concept of Node that becomes the element that hosts one or multiple pollers.
- Each node can host multiple polling instances in order to respond with more flexibility for the partitioning of the monitoring instances (for hosters, outsourcers etc…)
- Removal of experimental, non-viable and irelevent parameters from the legacy Nagios code in order to enhance performance and stability of the engine
- Replacement of obsolete libraries (PHP and JS)
- Full integration of the on the fly configuration reloading system: https://blog.centreon.com/centreon-engine-1-4-la-recharge-a-chaud-de-votre-configuration/?lang=fr
What’s new regarding the features ?
- The “real time” console of Centreon BAM will be included in the community version of Centreon in order to display business views and applications. Centreon Broker will directly calculate the values of Business Activities, performance will then be significantly improved in order to handle a greater number of Business Activities and Key Performance Indicators (KPI).
- Introduction of incident view : it’s a new view that aims to analyze impact on alerts, some root causes (RCA) and to follow incidents directly on the dashboard. This would allow us to view ticketing data on the interface.
- Data structure : the R&D team would like to offer new dimensions by adding new objects such as environments, domains and tags.
- Concept of organization to better split data : request from outsourcers that want to pool their Centreon in order to organize several clients on the same instance. Now, it will be much easier !
- The notion of severity is redesigned in order to exclude linear criticality. The severity will be used for emergency and impact purposes. It’s Centreon Broker that manages it as well in order to calculate in real time.
- A Rest API has been implemented for the management of all configuration objects and parameters in Centreon. Some data and parameters imports could be done from the CMDB or other reference sources.
- Redesign of notification management : the notification will be managed by Centreon Broker thus allowing to add a more performant system of correlation and to filter alerts. The notifications will be configurable by non-admin “user” profiles so that everyone can adjust himself his needs and desires in terms of notification.
Conclusion about Centreon
Many new features were announced for both the user interface and the monitoring engine.
Merethis leads Centreon to :
- a better scalability and stability
- a better monitoring software with better life cycle management of monitored resources
- a product for better analysis of performance and accessibility to help the IT technical teams to make the best possible choices.
Feedback from CIO
Fabrice Noel, CIO of the company Sodexi, gave us the honor of bringing his vision and expectations from a monitoring tool. Fabrice Noel is the CTO of a company of 400 people operating in the express delivery market. They have a 24/7 and worldwide business. Their clients are FedEx, DHL, TNT, La Poste, Chronopost, Colisposte, Air France / KLM …
In this context, the monitoring of their systems (in the broad sense) and the flow of information is the foundation of their quality services. Indeed, they have implemented a solution on the market since 2004 and there are only two people to share the constraint for fifty servers (50% Linux / 50% Win) and more than 300 “business “streams to monitor …
In the end and above all, what are the expectations of CIOs like Mr. Noel ?
Well, stability! A powerful tool with features is certainly important but “unstable monitoring tool goes directly to the trash!”
It is also important to remember that we do not monitor only machines but businesses!
These presentations lead to multiple exchanges among our guests and speakers.
The Centreon book giveaway
The Editions ENI kindly proposed to win the prize “Master the monitoring of your information system.” (French book : “Maitrisez la supervision de votre système d’information”.)
The winner was Leon, a Zabbix user. Congratulations !
The afternoon began with a demo of Centreon 3.
First conference : case study Centreon
Etienne Simonneau gave us a feedback on the use of Centreon in the IT of his company: Auchan.
From a community-based version Nagios and NDOutils, Auchan encountered performance issues due to an exceedingly high number of monitored resources: 80,000 services. Auchan with the help of Merethis decided to migrate towards Centreon Engine and Centreon Broker.
The result: the platform is now much lighter and Auchan is able to add even more indicators to continue to strengthen controls of its infrastructure.
Second conference : Centreon / Rudder, a winning team.
Nicolas Charles from the company Normation presented the configuration management tool, Rudder, developed by Normation and especially how Centreon and Rudder can be complementary.
Nicolas showed that after some joint work between Normation and Merethis, many things could be automated. Nicolas demonstrated that it is possible to automatically deploy management agents based on the attributes configured in Rudder. During the agent deployment, Rudder can communicate through the Centreon API to automatically add resources. In short, it saves a lot of time for the admins.
Third conference : Presentation of monitoring stack at Criteo
At last, Charles Judith accompanied by his colleague Hervé Chrun presented how Centreon is integrated with other tools such as Chef, RackMonkey, Gitorious, ModGearman in the Criteo company for the monitoring of all their servers.
This interesting presentation highlighted that application managers could very well handle the configuration of their indicators. It means that the configuration could be either delegated to server managers or it could be automated in order to cover everything.
At the end of the event, the participants and the Centreon / Merethis teams had pleasant discussions.
BarCamp # 5 attracted more “Merethisiens” to exchange around the Centreon Project:
Romain Le Merlus (one of the founding fathers of Centreon), Quentin Garnier (Architect Merethis), Rémi Werquin (QA manager), Etienne Gautier (Reporting Product Manager), Lionel Assepo, Alexandre Fouillé and even our interns (Julien and Maxime ;))
Thank you again to all participants for your visit, your mood and your multiple feedbacks!
See you in 2015 for the next BarCamp # 6 !!!
PS : some French feedbacks from our participants
- Matthieu Robin : http://www.matthieurobin.com/retour-du-centreon-barcamp-5/
- Monitoring-fr : http://www.monitoring-fr.org/2014/12/a-venir-sur-centreon/