Knowledge

• List vCenter default connectivity alarms
  Going to do this one the easy way. VMware publishes a table in their Online Library (see blow) that shows all the default. Below is a selection that excludes the performance alarms (see 8.7 for those).

  Alarm Name	Description
  Cannot Connect to Network	Monitors network connectivity on a vSwitch.
  Cannot Connect to Storage	Monitors host connectivity to a storage device.
  Cluster High Availability Error	Monitors high availability errors on a cluster.
  Datastore Usage On Disk	Monitors datastore disk usage.
  Exit Standby Error	Monitors whether a host cannot exit standby mode.
  Health Status Changed	Monitors changes to service and extension health status.
  Host Battery Status	Monitors host batteries.
  Host Connection and Power State	Monitors host connection and power state.
  Host Connection Failure	Monitors host connection failures.
  Host CPU Usage	Monitors host CPU usage.
  Host Error	Monitors host error and warning events.
  Host Hardware Fan Status	Monitors host fans.
  Host Hardware Power Status	Monitors host power.
  Host Hardware System Board Status	Monitors host system boards.
  Host Hardware Temperature Status	Monitors host temperature.
  Host Hardware Voltage	Monitors host voltage.
  Host Memory Status	Monitors host memory.
  Host Memory Usage	Monitors host memory usage.
  Host Processor Status	Monitors host processors.
  Host Service Console SwapIn Rate	Monitors host service console memory swapin rate.
  Host Service Console SwapOut Rate	Monitors host service console memory swapout rate.
  Host Status for Hardware Objects	Monitors the status of host hardware objects.
  Host Storage Status	Monitors host connectivity to storage devices.
  License Error	Monitors license errors.
  License Inventory Monitoring	Monitors the license inventory for compliancy.
  Migration Error	Monitors whether a virtual machine cannot migrate or relocate, or is orphaned.

• List possible actions for connectivity alarms
  Below you will find all actions that can be defind as a action for a specific alarm. (found here)

  Action	Description	Alarm Object
  Send a notification email	SMTP sends an email message. The SMTP must be ready when the email message is sent. You can set SMTP through vCenter Server or through Microsoft Outlook Express.	datacenter, datastore, cluster, host, resource pool, virtual machine, network, vNetwork distributed switch, dvPort group
  Send a notification trap	SNMP sends a notification trap. vCenter Server is the default SNMP notification receiver. An SNMP trap viewer is required to view a sent trap.	datacenter, datastore, cluster, host, resource pool, virtual machine
  Run a command	Performs the operation defined in the script you specify. It runs as separate process and does not block vCenter Server processes.	datacenter, datastore, cluster, host, resource pool, virtual machine, network, vNetwork distributed switch, dvPort group
  Enter or exit maintenance mode	Puts the host in and out of maintenance mode. Maintenance mode restricts virtual machine operations on the host. You put a host in maintenance mode when you need to move or service it.	host
  Enter or exit standby	Suspends or resumes the guest operating system on the virtual machine.	host
  Reboot or shut down host	Reboots or shuts down the host.	host

• For a given alarm, analyze and evaluate the affected virtual infrastructure components
  See the second column from the table above.
• Create a vCenter connectivity alarm
  To create a custom alarm you first need te specify the level you would like to create an alarm on. The Virtual Infrastrcture level is a good start (highest level). Do:
  • Select the tab “Alarms”
  • Select the view “Definitions”
  • Right click in the definitions pane and select “New Alarm…” the flowing pane opens:
    
    
  • Select the second bullet: “Monitor for specific events…”
  • Enter all values in all tabs to create the alarm. Since there is a large number of settings to be made I will not state all options. The options a quite clear as well.
• Relate the alarm to the affected components
  I think that this means that you can create an alarm on different levels like data-center, host or VM level. Actions that are triggered based on a event will only affect that level and the objects that are part of that level.
  
  For example, you can create a alarm for one VM only that monitors the VM and reboots the host when memory usage is out of control.
  
  Another answer to this question is perhaps how to see what part of your virtual infrastructure is affected by the issue indicated by the alarm. You can see this under “object” in the “Triggered Alarms” view.
  
  But perhaps something different is meant here. If you think so, help me out and reply below!

Tools

• vSphere Basic System Administration Guide
• Product Documentation
• vSphere Client