Design and deploy monitoring solution

[vincentnam]

Monitoring of any network or system is a crucial service for sustainable, maintenable and evolutive architecture. As it is a complex architecture, several level of monitoring are needed :

• low level : network monitoring for network security, prevent attacks and make it possible to monitor network load and size ressources
• mid level : service monitoring and track which service is used, which user do what to make it possible to easily administrate the architecture
• high level : which data is often requested, which data is never used to optimize data management and storage Network security should also be considered for the insurance of the communication into the architecture.

This part is mainly a "in-production" state service and the most of the work has to be done but there is already a list of tools that could be used :

• Openstack Ceilometer (mid level)
• Prometheus (low level)
• SNMP (low level) with Zabbix (low / mid level)
• Kerberos server (low level)
• DMZ

As the datalake is currently hosted on OSIRIM on virtual machine (on a VMWare platform), network architecture is a black box. Plateform-related tools could be selected for differents host platform.

There is still a lot of work to do on this part.

• [ ] Define needs for monitoring solutions
• [ ] State of art of solutions
• [ ] Add informations in documentation