Amazon CloudWatch
User Guide

Metrics Collected by the CloudWatch Agent

You can collect metrics from servers by installing the CloudWatch agent on the server. You can install the agent on both Amazon EC2 instances and on-premises servers, and on servers running either Linux or Windows Server. If you install the agent on an Amazon EC2 instance, the metrics it collects are in addition to the metrics enabled by default on Amazon EC2 instances.

For information about installing the CloudWatch agent on an instance, see Collect Metrics and Logs from Amazon EC2 Instances and On-Premises Servers with the CloudWatch Agent.

Metrics Collected by the CloudWatch Agent on Windows Server Instances

On a server running Windows Server, installing the CloudWatch agent enables you to collect the metrics associated with the counters in Windows Performance Monitor. The CloudWatch metric names for these counters are created by putting a space between the object name and the counter name. For example, the % Interrupt Time counter of the Processor object is given the metric name Processor % Interrupt Time in CloudWatch. For more information about Windows Performance Monitor counters, see the Microsoft Windows Server documentation.

The default namespace for metrics collected by the CloudWatch agent is CWAgent, although you can specify a different namespace when you configure the agent.

Metrics Collected by the CloudWatch Agent on Linux Instances

The metrics that you can collect with the CloudWatch agent on Linux instances are listed in the following table.

Metric Description

cpu_time_active

The amount of time that the CPU is active in any capacity. This metric is measured in hundredths of a second.

Unit: None

cpu_time_guest

The amount of time that the CPU is running a virtual CPU for a guest operating system. This metric is measured in hundredths of a second.

Unit: None

cpu_time_guest_nice

The amount of time that the CPU is running a virtual CPU for a guest operating system, which is low-priority and can be interrupted by other processes. This metric is measured in hundredths of a second.

Unit: None

cpu_time_idle

The amount of time that the CPU is idle. This metric is measured in hundredths of a second.

Unit: None

cpu_time_iowait

The amount of time that the CPU is waiting for I/O operations to complete. This metric is measured in hundredths of a second.

Unit: None

cpu_time_irq

The amount of time that the CPU is servicing interrupts. This metric is measured in hundredths of a second.

Unit: None

cpu_time_nice

The amount of time that the CPU is in user mode with low-priority processes, which can easily be interrupted by higher-priority processes. This metric is measured in hundredths of a second.

Unit: None

cpu_time_softirq

The amount of time that the CPU is servicing software interrupts. This metric is measured in hundredths of a second.

Unit: None

cpu_time_steal

The amount of time that the CPU is in stolen time, which is time spent in other operating systems in a virtualized environment. This metric is measured in hundredths of a second.

Unit: None

cpu_time_system

The amount of time that the CPU is in system mode. This metric is measured in hundredths of a second.

Unit: None

cpu_time_user

The amount of time that the CPU is in user mode. This metric is measured in hundredths of a second.

Unit: None

cpu_usage_active

The percentage of time that the CPU is active in any capacity.

Unit: Percent

cpu_usage_guest

The percentage of time that the CPU is running a virtual CPU for a guest operating system.

Unit: Percent

cpu_usage_guest_nice

The percentage of time that the CPU is running a virtual CPU for a guest operating system, which is low-priority and can be interrupted by other processes.

Unit: Percent

cpu_usage_idle

The percentage of time that the CPU is idle.

Unit: Percent

cpu_usage_iowait

The percentage of time that the CPU is waiting for I/O operations to complete.

Unit: Percent

cpu_usage_irq

The percentage of time that the CPU is servicing interrupts.

Unit: Percent

cpu_usage_nice

The percentage of time that the CPU is in user mode with low-priority processes, which can easily be interrupted by higher-priority processes.

Unit: Percent

cpu_usage_softirq

The percentage of time that the CPU is servicing software interrupts.

Unit: Percent

cpu_usage_steal

The percentage of time that the CPU is in stolen time, which is time spent in other operating systems in a virtualized environment.

Unit: Percent

cpu_usage_system

The percentage of time that the CPU is in system mode.

Unit: Percent

cpu_usage_user

The percentage of time that the CPU is in user mode.

Unit: Percent

disk_free

Free space on the disks.

Unit: Bytes

disk_inodes_free

The number of available index nodes on the disk.

Unit: Count

disk_inodes_total

The total number of index nodes reserved on the disk.

Unit: Count

disk_inodes_used

The number of used index nodes on the disk.

Unit: Count

disk_total

Total space on the disks, including used and free.

Unit: Bytes

disk_used

Used space on the disks.

Unit: Bytes

disk_used_percent

The percentage of total disk space that is used.

Unit: Percent

diskio_iops_in_progress

The number of I/O requests that have been issued to the device driver but have not yet completed.

Unit: Count

diskio_io_time

The amount of time that the disk has had I/O requests queued.

Unit: Milliseconds

diskio_reads

The number of disk read operations.

Unit: Count

diskio_read_bytes

The number of bytes read from the disks.

Unit: Bytes

diskio_read_time

The amount of time that read requests have waited on the disks. Multiple read requests waiting at the same time all increase the number. For example, if 5 requests all wait for an average of 100 milliseconds, then 500 is reported.

Unit: Milliseconds

diskio_writes

The number disk write operations.

Unit: Count

diskio_write_bytes

The number of bytes written to the disks.

Unit: Bytes

diskio_write_time

The amount of time that write requests have waited on the disks. Multiple write requests waiting at the same time all increase the number. For example, if 8 requests all wait for an average of 1000 milliseconds, then 8000 is reported.

Unit: Milliseconds

mem_active

The amount of memory that has been used in some way during the last sample period.

Unit: Bytes

mem_available

The amount of memory that is available and can be given instantly to processes.

Unit: Bytes

mem_available_percent

The percentage of memory that is available and can be given instantly to processes.

Unit: Percent

mem_buffered

The amount of memory that is being used for buffers.

Unit: Bytes

mem_cached

The amount of memory that is being used for file caches.

Unit: Bytes

mem_free

The amount of memory that is not being used.

Unit: Bytes

mem_inactive

The amount of memory that has not been used in some way during the last sample period

Unit: Bytes

mem_total

The total amount of memory.

Unit: Bytes

mem_used

The amount of memory currently in use.

Unit: Bytes

mem_used_percent

The percentage of memory currently in use.

Unit: Percent

net_bytes_recv

The number of bytes received by the network interface.

Unit: Bytes

net_bytes_sent

The number of bytes sent by the network interface.

Unit: Bytes

net_drop_in

The number of packets received by this network interface that were dropped.

Unit: Count

net_drop_out

The number of packets transmitted by this network interface that were dropped.

Unit: Count

net_err_in

The number of receive errors detected by this network interface.

Unit: Count

net_err_out

The number of transmit errors detected by this network interface.

Unit: Count

net_packets_sent

The number of packets sent by this network interface.

Unit: Count

net_packets_recv

The number of packets received by this network interface.

Unit: Count

netstat_tcp_close

The number of TCP connections with no state.

Unit: Count

netstat_tcp_close_wait

The number of TCP connections waiting for a termination request from the client.

Unit: Count

netstat_tcp_closing

The number of TCP connections that are waiting for a termination request with acknowledgement from the client.

Unit: Count

netstat_tcp_established

The number of TCP connections established.

Unit: Count

netstat_tcp_fin_wait1

The number of TCP connections in the FIN_WAIT1 state, during the process of closing a connection.

Unit: Count

netstat_tcp_fin_wait2

The number of TCP connections in the FIN_WAIT2 state, during the process of closing a connection.

Unit: Count

netstat_tcp_last_ack

The number of TCP connections waiting for the client to send acknowledgement of the connection termination message. This is the last state right before the connection is closed down.

Unit: Count

netstat_tcp_listen

The number of TCP ports currently listening for a connection request.

Unit: Count

netstat_tcp_none

The number of TCP connections with inactive clients.

Unit: Count

netstat_tcp_syn_sent

The number of TCP connections waiting for a matching connection request after having sent a connection request.

Unit: Count

netstat_tcp_syn_recv

The number of TCP connections waiting for connection request acknowledgement after having sent and received a connection request.

Unit: Count

netstat_tcp_time_wait

The number of TCP connections currently waiting to ensure that the client received the acknowledgement of its connection termination request.

Unit: Count

netstat_udp_socket

The number of current UDP connections.

Unit: Count

processes_blocked

The number of processes that are blocked.

Unit: Count

processes_dead

The number of processes that are "dead," which is indicated by the X state code on Linux.

Unit: Count

processes_idle

The number of processes that are idle (sleeping for more than 20 seconds). Available only on FreeBSD instances.

Unit: Count

processes_paging

The number of processes that are paging, which is indicated by the W state code on Linux.

Unit: Count

processes_running

The number of processes that are running, indicated by the R state code.

Unit: Count

processes_sleeping

The number of processes that are sleeping, indicated by the S state code.

Unit: Count

processes_stopped

The number of processes that are stopped, indicated by the T state code.

Unit: Count

processes_total

The total number of processes on the instance.

Unit: Count

processes_total_threads

The total number of threads making up the processes. This metric is available only on Linux instances.

Unit: Count

processes_wait

The number of processes that are paging, which is indicated by the W state code on FreeBSD instances. This metric is available only on FreeBSD instances.

Unit: Count

processes_zombies

The number of zombie processes, indicated by the Z state code.

Unit: Count

swap_free

The amount of swap space that is not being used.

Unit: Bytes

swap_used

The amount of swap space currently in use.

Unit: Bytes

swap_used_percent

The percentage of swap space currently in use.

Unit: Percent