Metrics collected by Lambda Insights

Lambda Insights collects several metrics from the Lambda functions where it is installed. Some of these metrics are available as time series aggregated data in CloudWatch Metrics. Other metrics are not aggregated into time series data but can be found in the embedded metric format log entries by using CloudWatch Logs Insights.

The following metrics are available as time series aggregated data in CloudWatch Metrics in the LambdaInsights namespace.

Metric name	Dimensions	Description
cpu_total_time	function_name function_name, version	Sum of cpu_system_time and cpu_user_time. Unit: Milliseconds
init_duration	function_name function_name, version	The amount of time spent in the init phase of the Lambda execution environment lifecycle. Unit: Milliseconds
memory_utilization	function_name function_name, version	The maximum memory measured as a percentage of the memory allocated to the function. Unit: Percent
rx_bytes	function_name function_name, version	The number of bytes received by the function. Unit: Bytes
tmp_used		The amount of space used in the /tmp directory. Unit: Bytes
tx_bytes	function_name function_name, version	The number of bytes sent by the function. Unit: Bytes
total_memory	function_name function_name, version	The amount of memory allocated to your Lambda function. This is the same as your function’s memory size. Unit: Megabytes
total_network	function_name function_name, version	Sum of rx_bytes and tx_bytes. Even for functions that don't perform I/O tasks, this value is usually greater than zero because of network calls made by the Lambda runtime. Unit: Bytes
used_memory_max	function_name function_name, version	The measured memory of the function sandbox. Unit: Megabytes

The following metrics can be found in the embedded metric format log entries by using CloudWatch Logs Insights. For more information about CloudWatch Logs Insights, see Analyzing Log Data with CloudWatch Logs Insights.

For more information about embedded metric format, see Embedding metrics within logs.

Metric name	Description
cpu_system_time	The amount of time the CPU spent executing kernel code. Unit: Milliseconds
cpu_total_time	Sum of cpu_system_time and cpu_user_time. Unit: Milliseconds
cpu_user_time	The amount of time the CPU spent executing user code. Unit: Milliseconds
fd_max	The maximum number of file descriptors available. Unit: Count
fd_use	The maximum number of file descriptors in use. Unit: Count
memory_utilization	The maximum memory measured as a percentage of the memory allocated to the function. Unit: Percent
rx_bytes	The number of bytes received by the function. Unit: Bytes
tx_bytes	The number of bytes sent by the function. Unit: Bytes
threads_max	The number of threads in use by the function process. As a function author, you don't control the initial number of threads created by the runtime. Unit: Count
tmp_max	The amount of space available in the /tmp directory. Unit: Bytes
total_memory	The amount of memory allocated to your Lambda function. This is the same as your function’s memory size. Unit: Megabytes
total_network	Sum of rx_bytes and tx_bytes. Even for functions that don't perform I/O tasks, this value is usually greater than zero because of network calls made by the Lambda runtime. Unit: Bytes
used_memory_max	The measured memory of the function sandbox. Unit: Bytes