Monitoring Platform Implementation Plan

Written by: Tian Yi ([email protected])

Basic goals

? Monitoring effectiveness: Monitoring can correctly respond to the system, application running state, fault can be timely alarm, monitoring the important objects, can locate the problem, not everything, increase the burden and complexity.
? Platform availability: The monitoring platform itself may also be at risk of failure, so a set of availability assurance mechanisms is needed. The first is the hardware level to ensure that some hardware failure, monitoring services are not affected, second, in case of system crashes, can rebuild the platform and data recovery.
? Information security: Network isolation and user authorization.

Program Design

? Monitoring architecture
Monitoring the sub-server and monitored side, where the server side of the deployment, and the monitoring side according to the monitoring requirements, some monitoring items need to be installed on the monitored side of the plug-in, and some do not need. According to this requirement to classify, monitor host survival, service port or Web URL, do not need to install plug-ins, and such as monitoring system load, disk space usage, number of processes, etc., belongs to the type of host resource monitoring, you need to install plug-ins.

? Platform Availability Architecture
Two levels, one is hardware redundancy and the other is a backup recovery mechanism.
(1) Hardware redundancy: A single physical host, to ensure the power and hard disk redundancy and fault tolerance, reduce the probability of down machine. Also configure the same configured physical host as a backup or standby machine.
(2) using virtualization at the bottom of the application layer, monitoring the platform operation and virtual machine, and automatically to the virtual machine offsite (machine) backup. When the system crashes, rebuild the system and restore the data quickly with virtual machine backups.

Technology implementation

? Object Monitoring implementation
The monitoring platform installs on the virtual machine, installs the necessary various builds (the Web, the database, the PHP and so on), can carry on the various management operation through the Web interface. The surveillance system uses the industry's well-known nagios, proven, and does not have legal and licensing risks. On a system that needs to monitor host resources, install the plugin Nrpe with Nagios, discarding the complexity of SNMP, which increases security and reduces the cost of resources.
? Implementation of availability
The bottom layer Proxmox virtualization with open source private cloud tools, creating two virtual machines on it, one for installing Nagios monitoring, and the other for installing NFS shared disks. Two physical hosts, one to run the monitoring system, and another to share the NFS disk space, is attached to the monitoring system, to do the monitoring system of virtual machine automatic backup use.
? Monitoring System Overview
Main projects include: Hosts, host groups, services, service groups, contacts, contact groups, monitoring status, and so on.

Implementation steps

? Physical machine shelves, power on, connect network cable, boot, hard disk raid.
? Install the underlying virtualized environment and ensure network connectivity.
? Configure the virtual machine environment and create the required virtual machines.
? Install the virtual machine operating system.
? Deploy Nagios to monitor and deploy NFS shared storage.
? Installed by the monitoring side plug-in, configure and start the daemon.
? Add the monitoring item and make it effective.
? Human simulated failure of the monitored terminal, check whether the monitoring is effective (fault alarm, recovery)

Tools List

? Monitoring System centreon2.8 (packaged Nagios), downloaded from Centreon official, without authorization.
? Monitoring System plugin Nagios-plugin, download from Nagios official website.
? Monitor Remote plugin Nrpe (nagios remote plugin executor) and download it from the Nagios official website.
? Shared storage Nfs,linux comes with.
? Virtualization Management platform Proxmox, download from its official website.
? Other tools such as Apache, PHP, MySQL, etc., are downloaded from the Internet.

Implementation requirements

? Monitoring server can access the monitored side, without the need for third-party forwarding;
? The monitoring server can access the public network, so as to be able to download the required software and dependencies, or can not be installed;
? If need from the public network operation monitoring platform, preferably can allocate public network IP;
? The monitored side needs to open the TCP 5666 port, can not be blocked by the firewall;
? A few system temporary simulation failure occurred, the effectiveness of testing monitoring;
? A list of persons using this monitor is required for reasonable authorization;
? For ease of configuration and post-maintenance, it is strongly recommended to set up remote login permissions (such as * * * authorization), so as to greatly improve efficiency;
? Allocate enough IP addresses to use.

Precautions

? When deploying the monitoring server side, ensure that the IP settings are unique and do not conflict with the IP of other hosts;
? When the monitoring side deploys the Nrpe, it is necessary to obtain the consent of the relevant personnel; The firewall open port (if enabled) needs to be supported and then down, and cannot cause a business outage due to the installation.

Appendix: Server Configuration Requirements
Name Specification Model Quantity remark
Cpu Log strong 2600 series 2 pcs 2 machines total four PCs
Memory DDR4 16G 64G Total 8
Hard drive Sas 15000 rpm 600G at least 3 blocks total 6 or more
Power supply Dual power supply 1 4 x

More systematic and instantiated Proxmox hyper-converged private Cloud Practice series, please visit my column "Everyone can play the private cloud artifact-proxmox", Poke here, a moment to direct!

Monitoring Platform Implementation Plan