18. Monitoring VMware
by Andrea Vacche, Patrik Uytterhoeven, Rihards Olups
Zabbix: Enterprise Network Monitoring Made Easy
- Zabbix: Enterprise Network Monitoring Made Easy
- Table of Contents
- Zabbix: Enterprise Network Monitoring Made Easy
- Zabbix: Enterprise Network Monitoring Made Easy
- Credits
- Preface
- I. Module 1
- 1. Getting Started with Zabbix
- The first steps in monitoring
- Zabbix features and architecture
- Installation
- Summary
- 2. Getting Your First Notification
- 3. Monitoring with Zabbix Agents and Basic Protocols
- 4. Monitoring SNMP Devices
- 5. Managing Hosts, Users, and Permissions
- 6. Detecting Problems with Triggers
- 7. Acting upon Monitored Conditions
- 8. Simplifying Complex Configurations with Templates
- 9. Visualizing Data with Graphs and Maps
- 10. Visualizing Data with Screens and Slideshows
- 11. Advanced Item Monitoring
- 12. Automating Configuration
- 13. Monitoring Web Pages
- 14. Monitoring Windows
- 15. High-Level Business Service Monitoring
- 16. Monitoring IPMI Devices
- 17. Monitoring Java Applications
- 18. Monitoring VMware
- 19. Using Proxies to Monitor Remote Locations
- 20. Encrypting Zabbix Traffic
- 21. Working Closely with Data
- 22. Zabbix Maintenance
- A. Troubleshooting
- B. Being Part of the Community
- 1. Getting Started with Zabbix
- II. Module 2
- 1. Zabbix Configuration
- 2. Getting Around in Zabbix
- 3. Groups, Users, and Permissions
- 4. Monitoring with Zabbix
- Introduction
- Active agents
- Passive agents
- Extending agents
- SNMP checks
- Internal checks
- Zabbix trapper
- IPMI checks
- JMX checks
- Aggregate checks
- External checks
- Database monitoring
- Checks with SSH
- Checks with Telnet
- Calculated checks
- Building web scenarios
- Monitoring web scenarios
- Some advanced monitoring tricks
- Autoinventory
- 5. Testing with Triggers in Zabbix
- 6. Working with Templates
- 7. Data Visualization and Reporting in Zabbix
- 8. Monitoring VMware and Proxies
- 9. Autodiscovery
- 10. Zabbix Maintenance and API
- C. Upgrading and Troubleshooting Zabbix
- III. Module 3
- 1. Deploying Zabbix
- 2. Distributed Monitoring
- 3. High Availability and Failover
- Understanding high availability
- Understanding the levels of IT service
- Some considerations about high availability
- Implementing high availability on a web server
- Configuring the Zabbix server for high availability
- Implementing high availability for a database
- Summary
- 4. Collecting Data
- 5. Visualizing Data
- 6. Managing Alerts
- 7. Managing Templates
- 8. Handling External Scripts
- 9. Extending Zabbix
- 10. Integrating Zabbix
- D. Bibliography
- Index
There are a lot of virtualization solutions available today. Their target markets and popularity differ, but for enterprise shops that can afford it, VMware solutions are quite widespread. Zabbix offers built-in support for monitoring VMware. This support includes:
- Monitoring vSphere and/or vCenter
- Automatically discovering all hypervisors
- Automatically discovering all virtual machines
Monitoring VMware does not involve any custom layers—Zabbix accesses the VMware API directly, and the monitoring of such an environment is very easy to set up. For this chapter, you will need access to a VMware instance API, including a username and password. It might be a good idea to try this on a smaller or non-production environment first.
Note
If discovering a large environment from a vCenter, the vCenter API endpoint could get overloaded, as Zabbix would connect to it and request data for all the vSphere instances and virtual machines that have been discovered. It might make sense to split the monitoring over individual vSphere instances instead.
To try out VMware support, we will need:
- The IP address or hostname on which we have access to the VMware API
- The username for an account with permissions to retrieve the information
- The password for that account
First, the server must be compiled with VMware support. If you have installed from packages, the support most likely is included. If you installed from source, check whether the Zabbix server log file lists VMware monitoring as enabled:
VMware monitoring: YES
When compiling from source, the following options are needed for VMware support:
--with-libcurl--with-libxml2
As with several other features we have explored so far, the Zabbix server doesn't start any VMware-specific processes by default. Edit zabbix_server.conf and look for the StartVMwareCollectors parameter. Add a new line and tell Zabbix to start two VMware collectors:
StartVMwareCollectors=2
Restart the server. Why two collectors? Zabbix developers recommend the number of the collectors to be based on the number of monitored VMware instances. For the best performance, it is suggested to start more collectors than the monitored instance count, but less than double the monitored instance count. Or, if we put that in an equation, instances < StartVMwareCollectors < ( instances * 2 ). We start small and monitor a single instance for now, so we'll have 1 < StartVMwareCollectors < 2. It is also recommended to always start at least two VMware collectors, so the choice is obvious here. If we had two VMware instances to monitor, it would be three collectors: 2 < StartVMwareCollectors < 4.
We will start by unleashing Zabbix at the VMware API and allowing it to automatically discover everything using the templates that are shipped with Zabbix. Once we see that it works as expected, we will discover how we can customize and expand this monitoring, as well as looking under the hood a bit at the mechanics of VMware monitoring.
-
No Comment