Overview
The VMware service monitors allow you to monitor and alert based on the performance and status your virtual resources. These monitors can watch for threshold violations with computing resources for VMs, ESX servers, and changes in power states.
...
ESX Server Monitors
ESX Server monitors focus on the ESX server host, as a physical computing resource, for monitoring and alerting.
There are currently three ESX-related monitors:
ESX (Advanced Metrics): uses an up.time agent on the ESX server
vSphere ESX Server Performance: uses metrics transferred to up.time using vSync
ESX Workload monitor: a legacy monitor that can no longer be added to up.time, and is found only in upgraded up.time deployments
The metrics collected for these ESX server monitors can be used to trigger up.time alerts and actions. These performance monitors can answer questions such as the following:
Are CPU or memory usage on the host too high?
Are network and disk I/O usage or latency within acceptable limits?
Are disk and network error rates too high?
Are memory ballooning targets being exceeded?
vSphere ESX Server Performance
The vSphere ESX Server Performance monitor allows you to alert based on performance checks on ESX server Elements managed by VMware vSphere, but monitored in up.time via vSync.
vSphere ESX Server Performance Monitor Metrics
The following vSphere metric types for ESX server performance can be used to configure thresholds in up.time :
Time Interval | A positive integer indicating the number of minutes’ worth of performance data samples to average, then compare against threshold definitions (default: 30). | |
CPU Check: | Warning- and critical-level thresholds can be set, using positive integers, for average CPU usage as either percentage usage, or average MHz usage. | |
Memory Check: | Warning- and critical-level thresholds can be set, using positive integers, for one of the following value types: Usage (%) - percentage of total configured or available memory used Memory Consumed (MB) - amount of memory consumed by VMs on this host Memory Active (MB) - amount of memory actively used by VMs on this host Balloon Memory (MB) - amount of memory allocated by vmmemctl across all VMs on this host Zero Memory (KB) - memory that only contains 0s allocated to VMs | |
Swap Check: | Warning- and critical-level thresholds can be set, using positive integers, for either swap space used (in MB), or swap rate (the combined swap-in rate and swap-out rate, in KBps, across all VMs on this host). | |
Disk Device I/O: | Warning- and critical-level thresholds can be set, using positive integers, for one of the following value types: Usage (KBps) - aggregate disk I/O rate across all VMs on the host Physical Device Command Latency (ms) - average time to process a read and write from the physical device Queue Command Latency (ms) - average time spent in the VMkernel queue per SCSI command Command Latency (ms) - average time taken to process a SCSI command issued by the Guest OS to the VM Checks are made against the average for all detected disk devices, or any individual device that is violating the threshold. | |
Disk Device Errors Check: | Warning- and critical-level thresholds can be set, using positive integers, for either the number of SCSI command aborts per minute, or the number of bus resets per minute. Checks are made against the average for all detected disk devices, or any individual device that is violating the threshold. | |
Network I/O: | Warning- and critical-level thresholds can be set, using positive integers, for the aggregate received and transmitted rate (in KBps). Checks are made against the average for all detected network interfaces, or any individual network interface that is violating the threshold. | |
Network Errors Check: | Warning- and critical-level thresholds can be set, using positive integers, for the aggregate received and transmitted packets dropped per minute. Checks are made against the average for all detected network interfaces, or any individual network interface that is violating the threshold. |
Configuring vSphere ESX Server Performance Monitors
To configure a vSphere ESX Server Performance monitor, do the following:
Select the monitor from the Add Service Monitor window, in the VMware Monitors section.
Click Continue to begin configuring the service monitor.
Complete the monitor information fields.
See Monitor Identification for more information on configuring service monitor information fields.
Info |
---|
When selecting an Element associated with this service monitor, only ESX servers monitored in up.time via vSync will appear in the Single System list |
In the vSphere ESX Server Performance Settings section, in the Time Interval sub-section, enter the number of minutes’ worth of time samples that will be used to compare thresholds.
For the following metric categories, select the metric unit of measurement, then configure the monitor’s warning- or critical-level threshold values:
- CPU Usage
- Memory
- Swap
- Disk Device I/O
- Disk Errors
- Network I/O
- Network Errors
For more information on these metrics, see VMware Monitors.
For more information about setting thresholds, see Configuring Warning and Critical Thresholds.
Complete the following settings:
- Timing Settings (see Adding Monitor Timing Settings Information for more information)
- Alert Settings (see Monitor Alert Settings for more information)
- Monitoring Period settings (see Monitor Timing Settings for more information)
- Alert Profile settings (see Alert Profiles for more information)
- Action Profile settings (see Action Profiles for more information)
Click Finish.
ESX (Advanced Metrics)
The ESX (Advanced Metrics) monitor offers greater visibility into your ESX environment by expanding on the high level usage metrics for a virtual machine’s CPU, memory, and disk activity.
ESX Advanced Metrics Monitor Metrics
The following ESX server metrics can be used to configure thresholds:
Percent Wait | Guest metric - The percentage of time that a virtual CPU was not running. A non-running CPU could be idle (halted) or waiting for an external event such as I/O. | |
Memory Balloon (Avg) | Guest metric - The average amount of memory, in KB, held by memory control for ballooning. | |
Memory Balloon Target | Guest metric - The total amount of memory, in KB, that can be used by memory control for ballooning. | |
Memory Overhead (Avg) | Guest metric - The average amount of additional host memory, in KB, allocated to the virtual machine. | |
Memory Swap In (Avg) | Guest metric - The average amount of memory, in KB, that was swapped in. | |
Memory Swap Out (Avg) | Guest metric - The average amount of memory, in KB, that was swapped out. | |
Memory Zero (Avg) | Guest metric - The average amount of memory, in KB, that was zeroed out. | |
Memory Swap Used (Avg) | Host metric - The average amount of memory, in KB, that was used by the swap file. | |
Memory Swap Target | Guest metric - The total amount of memory, in KB, that can be swapped. | |
Disk Total Latency | Host metric - The average time, in milliseconds, taken for disk commands by a guest OS. This is the sum of kernelCommandLatency and | |
Disk Kernel Latency | Host metric - The average time, in milliseconds, spent in the ESX Server VMkernel per command. | |
Disk Device Latency | Host metric - The average time, in milliseconds, taken to complete a command from the physical device. | |
Disk Queue Latency | Host metric - The average time, in milliseconds, spent in the ESX Server VMkernel queue per write. | |
Disk Commands Aborted | Host metric - The number of disk commands aborted during the defined interval. | |
Disk Commands Issued | Host metric - The number of disk commands issued during the defined interval. | |
Disk Bus Resets | Host metric - The number of bus resets during the defined interval. |
Configuring ESX (Advanced Metrics) Monitors
To configure an ESX (Advanced Metrics) monitor, do the following:
Select the monitor from the Add Service Monitor window, in the VMware Monitors section.
Click Continue to begin configuring the service monitor.
Complete the monitor information fields.
See Monitor Identification for more information on configuring service monitor information fields.
In the ESX (Advanced Metrics) Settings section, configure the monitor’s warning- and critical-level alerting thresholds by completing the following fields:
- Percent Wait
- Memory Balloon
- Memory Balloon Target
- Memory Overhead
- Memory Swap In
- Memory Swap Out
- Memory Zero
- Memory Swap Used
- Memory Swap Target
- Disk Total Latency
- Disk Kernel Latency
- Disk Device Latency
- Disk Queue Latency
- Disk Commands Aborted
- Disk Commands Issued
- Disk Bus Resets
- Response time
For more information on these metrics, see VMware Monitors.
For more information about setting thresholds and response time, see Configuring Warning and Critical Thresholds.
Complete the following settings:
- Timing Settings (see Adding Monitor Timing Settings Information for more information)
- Alert Settings (see Monitor Alert Settings for more information)
- Monitoring Period settings (see Monitor Timing Settings for more information)
- Alert Profile settings (see Alert Profiles for more information)
- Action Profile settings (see Action Profiles for more information)
Click Finish.
ESX Workload
The ESX Workload monitor collects a set of metrics from all of the instances that are running on an ESX v3 or v4 server over a specified time period.
Info |
---|
This monitor is a legacy monitor that cannot be added to your up.time configuration as a new service monitor; it exists in upgraded configurations that originally included it, and works only with the VMware ESX type Element. |
The monitor the compares the highest values returned by the instances and then compares them to the thresholds that you set. If the values exceed the thresholds, up.time issues an alert. The monitor does not pinpoint the specific instance(s) that have exceeded the defined thresholds.
For example, you are monitoring an ESX server that is running three instances. You configured the ESX Workload monitor to collect data samples every 10 minutes, and to issue a warning when memory usage exceeds 300 MB. The three instances are using the following amounts of memory: 110 MB, 227 MB, and 315 MB. The ESX Workload monitor focuses on the value of 315 MB and, since it exceeds the warning threshold, issues an alert.
ESX Workload Monitor Metrics
The following metrics are used by the ESX Workload monitor:
Time Interval | The amount of time, in minutes, at which the monitor will collect data samples from the ESX server. | |
CPU Warning Threshold | The amount of CPU resources, measured in megahertz (MHz), that the instances on the ESX server must consume before up.time issues a warning. | |
CPU Critical Threshold | The amount of CPU resources, measured in megahertz MHz, that the instances on the ESX server must consume before up.time issues a critical alert. | |
Network Bandwidth Warning Threshold | The amount of network traffic in and out of the server, measured in megabits per second (Mbit/s), that must be exceeded before up.time issues a warning. | |
Network Bandwidth Critical Threshold | The amount of network traffic in and out of the server, measured in megabits per second (Mbit/s), that must be exceeded before up.time issues a critical alert. | |
Disk Usage Warning Threshold | The amount of data being written to the server’s hard disk, measured in kilobytes per second (KB/s), that must be exceeded before up.time issues a warning. | |
Disk Usage Critical Threshold | The amount of data being written to the server’s hard disk, measured in kilobytes per second (KB/s), that must be exceeded before up.time issues a critical alert. | |
Memory Usage Warning Threshold | The amount of overall system memory, measured in megabytes (MB), that must be exceeded before up.time issues a warning. | |
Memory Usage Critical Threshold | The amount of overall system memory, measured in megabytes (MB), that must be exceeded before up.time issues a critical alert. | |
Percent Ready Warning Threshold | The percentage of time that one or more instances running on an ESX server is ready to run, but cannot run because it cannot access the processor on the ESX server. If the valued returned from the server exceeds this threshold, then up.time issues a warning. | |
Percent Ready Critical Threshold | The percentage of time that one or more instances running on an ESX server is ready to run, but cannot run because it cannot access the processor on the ESX server. If the valued returned from the server exceeds this threshold, then up.time issues a critical alert. | |
Percent Used Warning Threshold | The percentage of CPU time that an instance running on an ESX server is using. If the valued returned from the server exceeds this threshold, then up.time issues a warning. | |
Percent Used Critical Threshold | The percentage of CPU time that an instance running on an ESX server is using. If the valued returned from the server exceeds this threshold, then up.time issues a critical alert. |
Modifying an ESX Workload Monitor Configuration
To modify the configuration of a legacy ESX Workload monitor, do the following:
If required, change the monitor information fields.
See Monitor Identification for more information.
In the ESX Workload Settings section, modify any of the monitor’s existing warning- or critical-level threshold values:
- Time Interval
- CPU Usage
- Network Bandwidth Usage
- Disk Usage
- Memory Usage
- Percent Ready
- Percent Used
For more information on these metrics, see VMware Monitors.
For more information about setting thresholds, see Configuring Warning and Critical Thresholds.
Complete the following settings:
- Timing Settings (see Adding Monitor Timing Settings Information for more information)
- Alert Settings (see Monitor Alert Settings for more information)
- Monitoring Period settings (see Monitor Timing Settings for more information)
- Alert Profile settings (see Alert Profiles for more information)
- Action Profile settings (see Action Profiles for more information)
Click Finish.
...