GetMetricData
You can use the GetMetricData
API to retrieve CloudWatch metric
values. The operation can also include a CloudWatch Metrics Insights query, and
one or more metric math functions.
A GetMetricData
operation that does not include a query can retrieve
as many as 500 different metrics in a single request, with a total of as many as 100,800
data points. You can also optionally perform metric math expressions on the values of
the returned statistics, to create new time series that represent new insights into your
data. For example, using Lambda metrics, you could divide the Errors metric by the
Invocations metric to get an error rate time series. For more information about metric
math expressions, see Metric Math Syntax and Functions in the Amazon CloudWatch User
Guide.
If you include a Metrics Insights query, each GetMetricData
operation can
include only one query. But the same GetMetricData
operation can also
retrieve other metrics. Metrics Insights queries can query only the most recent three
hours of metric data. For more information about Metrics Insights, see Query your metrics with CloudWatch Metrics Insights.
Calls to the GetMetricData
API have a different pricing structure than
calls to GetMetricStatistics
. For more information about pricing, see
Amazon CloudWatch
Pricing
Amazon CloudWatch retains metric data as follows:
-
Data points with a period of less than 60 seconds are available for 3 hours. These data points are high-resolution metrics and are available only for custom metrics that have been defined with a
StorageResolution
of 1. -
Data points with a period of 60 seconds (1-minute) are available for 15 days.
-
Data points with a period of 300 seconds (5-minute) are available for 63 days.
-
Data points with a period of 3600 seconds (1 hour) are available for 455 days (15 months).
Data points that are initially published with a shorter period are aggregated together for long-term storage. For example, if you collect data using a period of 1 minute, the data remains available for 15 days with 1-minute resolution. After 15 days, this data is still available, but is aggregated and retrievable only with a resolution of 5 minutes. After 63 days, the data is further aggregated and is available with a resolution of 1 hour.
If you omit Unit
in your request, all data that was collected with any
unit is returned, along with the corresponding units that were specified when the data
was reported to CloudWatch. If you specify a unit, the operation returns only data that
was collected with that unit specified. If you specify a unit that does not match the
data collected, the results of the operation are null. CloudWatch does not perform unit
conversions.
Using Metrics Insights queries with metric math
You can't mix a Metric Insights query and metric math syntax in the same expression, but you can reference results from a Metrics Insights query within other Metric math expressions. A Metrics Insights query without a GROUP BY clause returns a single time-series (TS), and can be used as input for a metric math expression that expects a single time series. A Metrics Insights query with a GROUP BY clause returns an array of time-series (TS[]), and can be used as input for a metric math expression that expects an array of time series.
Request Parameters
For information about the parameters that are common to all actions, see Common Parameters.
- EndTime
-
The time stamp indicating the latest data to be returned.
The value specified is exclusive; results include data points up to the specified time stamp.
For better performance, specify
StartTime
andEndTime
values that align with the value of the metric'sPeriod
and sync up with the beginning and end of an hour. For example, if thePeriod
of a metric is 5 minutes, specifying 12:05 or 12:30 asEndTime
can get a faster response from CloudWatch than setting 12:07 or 12:29 as theEndTime
.Type: Timestamp
Required: Yes
- LabelOptions
-
This structure includes the
Timezone
parameter, which you can use to specify your time zone so that the labels of returned data display the correct time for your time zone.Type: LabelOptions object
Required: No
- MaxDatapoints
-
The maximum number of data points the request should return before paginating. If you omit this, the default of 100,800 is used.
Type: Integer
Required: No
- MetricDataQueries.member.N
-
The metric queries to be returned. A single
GetMetricData
call can include as many as 500MetricDataQuery
structures. Each of these structures can specify either a metric to retrieve, a Metrics Insights query, or a math expression to perform on retrieved data.Type: Array of MetricDataQuery objects
Required: Yes
- NextToken
-
Include this value, if it was returned by the previous
GetMetricData
operation, to get the next set of data points.Type: String
Required: No
- ScanBy
-
The order in which data points should be returned.
TimestampDescending
returns the newest data first and paginates when theMaxDatapoints
limit is reached.TimestampAscending
returns the oldest data first and paginates when theMaxDatapoints
limit is reached.If you omit this parameter, the default of
TimestampDescending
is used.Type: String
Valid Values:
TimestampDescending | TimestampAscending
Required: No
- StartTime
-
The time stamp indicating the earliest data to be returned.
The value specified is inclusive; results include data points with the specified time stamp.
CloudWatch rounds the specified time stamp as follows:
-
Start time less than 15 days ago - Round down to the nearest whole minute. For example, 12:32:34 is rounded down to 12:32:00.
-
Start time between 15 and 63 days ago - Round down to the nearest 5-minute clock interval. For example, 12:32:34 is rounded down to 12:30:00.
-
Start time greater than 63 days ago - Round down to the nearest 1-hour clock interval. For example, 12:32:34 is rounded down to 12:00:00.
If you set
Period
to 5, 10, or 30, the start time of your request is rounded down to the nearest time that corresponds to even 5-, 10-, or 30-second divisions of a minute. For example, if you make a query at (HH:mm:ss) 01:05:23 for the previous 10-second period, the start time of your request is rounded down and you receive data from 01:05:10 to 01:05:20. If you make a query at 15:07:17 for the previous 5 minutes of data, using a period of 5 seconds, you receive data timestamped between 15:02:15 and 15:07:15.For better performance, specify
StartTime
andEndTime
values that align with the value of the metric'sPeriod
and sync up with the beginning and end of an hour. For example, if thePeriod
of a metric is 5 minutes, specifying 12:05 or 12:30 asStartTime
can get a faster response from CloudWatch than setting 12:07 or 12:29 as theStartTime
.Type: Timestamp
Required: Yes
-
Response Elements
The following elements are returned by the service.
- Messages.member.N
-
Contains a message about this
GetMetricData
operation, if the operation results in such a message. An example of a message that might be returned isMaximum number of allowed metrics exceeded
. If there is a message, as much of the operation as possible is still executed.A message appears here only if it is related to the global
GetMetricData
operation. Any message about a specific metric returned by the operation appears in theMetricDataResult
object returned for that metric.Type: Array of MessageData objects
- MetricDataResults.member.N
-
The metrics that are returned, including the metric name, namespace, and dimensions.
Type: Array of MetricDataResult objects
- NextToken
-
A token that marks the next batch of returned results.
Type: String
Errors
For information about the errors that are common to all actions, see Common Errors.
- InvalidNextToken
-
The next token specified is invalid.
HTTP Status Code: 400
Examples
Example
The following example requests a Metrics Insights query for aggregated
CPUUtilization
, and a metric.
Sample Request
{
"StartTime": 1637061900,
"EndTime": 1637074500,
"MetricDataQueries": [
{
"Expression": "SELECT AVG(CPUUtilization) FROM SCHEMA(\"AWS/EC2\", InstanceId)",
"Id": "q1",
"Period": 300,
"Label": "Cluster CpuUtilization"
},
{
"Id": "m1",
"Label": "Unhealthy Behind Load Balancer",
"MetricStat": {
"Metric": {
"Namespace": "AWS/ApplicationELB",
"MetricName": "UnHealthyHostCount",
"Dimensions": [
{
"Name": "TargetGroup",
"Value": "targetgroup/EC2Co-Defau-EXAMPLEWNAD/89cc68152b367e5f"
},
{
"Name": "LoadBalancer",
"Value": "app/EC2Co-EcsEl-EXAMPLE69Q/fdd2210e799e4376"
}
]
},
"Period": 300,
"Stat": "Average"
}
}
}
Sample Response
{
"Messages": [],
"MetricDataResults": [
{
"Id": "m1",
"Label": "Unhealthy Behind Load Balancer",
"StatusCode": "Complete",
"Timestamps": [
1637074200,
1637073900,
1637073600
],
"Values": [
0,
0,
0
]
},
{
"Id": "q1",
"Label": "Cluster CpuUtilization",
"StatusCode": "Complete",
"Timestamps": [
1637074245,
1637073945,
1637073645
],
"Values": [
1.2158469945359334,
0.8678863271635757,
0.7201860957623283
]
}
]
}
Example
The following example includes a Metrics Insights query for that is given the
ID error_rate
. The returned results of the query are then used in
the metric math expression to return availability
.
Sample Request
{
"StartTime": 1518867432,
"EndTime": 1518868032,
"MetricQueries": [
{
"Id": "availability",
"Expression": "(1 - error_rate) * 100",
"Label": "Availability"
},
{
"Id": " error_rate",
"Expression": "SELECT AVG(ErrorRate) FROM MyService",
"Period": 300,
"ReturnData": false
}
]
}
Example
The following example requests three separate metrics across two namespaces.
The labels of the first two metrics use dynamic labels to display the peak value
of CPUUtilization
during the time shown on the graph, and also the
time that the peak value was recorded. The Timezone
setting
specifies that the times shown in those dynamic labels reflect the United States
Eastern time zone, which is 4 hours behind UTC.
Sample Request
{
"StartTime": 1518867432,
"EndTime": 1518868232,
"LabelOptions": {
"Timezone" : "-0400"
},
"MetricDataQueries": [
{
"Id": "m1",
"Label": "CPUUtilization, peak of ${MAX} was at ${MAX_TIME}",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "CPUUtilization",
"Dimensions": [
{
"Name": "InstanceId",
"Value": "i-1234567890abcdef0"
}
]
},
"Period": 300,
"Stat": "Average"
}
},
{
"Id": "m2",
"Label": "CPUUtilization, peak of ${MAX} was at ${MAX_TIME}",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "CPUUtilization",
"Dimensions": [
{
"Name": "InstanceId",
"Value": "i-111111111111111111"
}
]
},
"Period": 300,
"Stat": "Average"
}
},
{
"Id": "m3",
"MetricStat": {
"Metric": {
"Namespace": "AWS/ELB",
"MetricName": "HealthyHostCount",
"Dimensions": [
{
"Name": "LoadBalancerName",
"Value": "my-lb-B"
}
]
},
"Period": 300,
"Stat": "Sum",
"Unit": "None"
}
}
]
}
Sample Response
{
"MetricDataResults": [
{
"Id": "m1",
"StatusCode": "Complete",
"Label": "CPUUtilization, peak of 31.5 was at 1-22 13:05",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
15000,
14000,
16000
]
},
{
"Id": "m2",
"StatusCode": "Complete",
"Label": "CPUUtilization, peak of 63.2 was at 1-22 13:20",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
15,
14,
16
]
},
{
"Id": "m3",
"StatusCode": "Complete",
"Label": "AWS/EC2 HealthyHostCount",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
15,
14,
16
]
}
]
}
Example
The following example retrieves the NetworkOut
metric for two
Auto Scaling groups, and uses them in an expression. These two metrics are
called m1 and m2, and the expression calculates e1 as the results of m2/m1. The
raw values and time stamps of the NetworkOut
metrics are not
returned.
Sample Request
{
"StartTime": 1518867432,
"EndTime": 1518868232,
"MetricQueries": [
{
"Id": "e1",
"Expression": "m2 / m1",
"Label": "my-asg-B / my-asg-A"
},
{
"Id": "m1",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "NetworkOut",
"Dimensions": [
{
"Name": "AutoScalingGroupName",
"Value": "my-asg-A"
}
]
},
"Period": 300,
"Stat": "SampleCount",
"Unit": "Bytes"
},
"ReturnData": false
},
{
"Id": "m2",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "NetworkOut",
"Dimensions": [
{
"Name": "AutoScalingGroupName",
"Value": "my-asg-B"
}
]
},
"Period": 300,
"Stat": "SampleCount",
"Unit": "Bytes"
},
"ReturnData": false
}
]
}
Sample Response
{
"MetricDataResults": [
{
"Id": "m1",
"StatusCode": "Complete"
},
{
"Id": "m2",
"StatusCode": "Complete"
},
{
"Id": "e1",
"StatusCode": "Complete",
"Label": "my-asg-B / my-asg-A",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
100,
100,
100
]
}
]
}
Example
In the following example, two levels of metric math expressions are used, with the result of one expression used as an input to the next expression:
Sample Request
{
"StartTime": 1518867432,
"EndTime": 1518868232,
"MetricDataQueries": [
{
"Id": "e1",
"Expression": "e2 + m3",
"Label": "my-asg-A * my-asg-B + my-asg-C"
},
{
"Id": "e2",
"Expression": "m1 * m2",
"Label": "my-asg-A * my-asg-B"
},
{
"Id": "m1",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "NetworkOut",
"Dimensions": [
{
"Name": "AutoScalingGroupName",
"Value": "my-asg-A"
}
]
},
"Period": 300,
"Stat": "SampleCount",
"Unit": "Bytes"
},
"ReturnData": false
},
{
"Id": "m2",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "NetworkOut",
"Dimensions": [
{
"Name": "AutoScalingGroupName",
"Value": "my-asg-B"
}
]
},
"Period": 300,
"Stat": "SampleCount",
"Unit": "Bytes"
},
"ReturnData": false
},
{
"Id": "m3",
"MetricStat": {
"Metric": {
"Namespace": "AWS/EC2",
"MetricName": "NetworkOut",
"Dimensions": [
{
"Name": "AutoScalingGroupName",
"Value": "my-asg-C"
}
]
},
"Period": 300,
"Stat": "SampleCount",
"Unit": "Bytes"
},
"ReturnData": false
}
]
}
Sample Response
{
"MetricDataResults": [
{
"Id": "m1",
"StatusCode": "Complete"
},
{
"Id": "m2",
"StatusCode": "Complete"
},
{
"Id": "m3",
"StatusCode": "Complete"
},
{
"Id": "e1",
"StatusCode": "Complete",
"Label": "my-asg-A * my-asg-B + my-asg-C",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
200,
200,
200
]
},
{
"Id": "e2",
"StatusCode": "Complete",
"Label": "my-asg-A * my-asg-B",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
100,
100,
100
]
}
]
}
Example
In the following example, custom metrics are searched and assigned IDs that
contain either "error" or "request", even if the original metric names did not
contain those words. Then an error rate is calculated using the
METRICS("string")
function on the assigned IDs.
Sample Request
{
"StartTime": 1518867432,
"EndTime": 1518868432,
"MetricDataQueries": [
{
"Id": "errorRate",
"Label": "Error Rate",
"Expression": "errors/requests"
},
{
"Id": "errorRatePercent",
"Label": "% Error Rate",
"Expression": "errorRate*100"
},
{
"Id": "requests",
"Expression": "SUM(METRICS('request'))",
"ReturnData": false
},
{
"Id": "errors",
"Expression": "SUM(METRICS('error'))",
"ReturnData": false
},
{
"Id": "error1",
"MetricStat": {
"Metric": {
"Namespace": "MyService",
"MetricName": "BadRequests",
"Dimensions": [
{
"Name": "Component",
"Value": "component-1"
}
]
},
"Period": 60,
"Stat": "Sum"
},
"ReturnData": false
},
{
"Id": "error2",
"MetricStat": {
"Metric": {
"Namespace": "MyService",
"MetricName": "ConnectionErrors",
"Dimensions": [
{
"Name": "Component",
"Value": "component-1"
}
]
},
"Period": 60,
"Stat": "Sum"
},
"ReturnData": false
},
{
"Id": "request1",
"MetricStat": {
"Metric": {
"Namespace": "MyService",
"MetricName": "InternalRequests",
"Dimensions": [
{
"Name": "Component",
"Value": "component-1"
}
]
},
"Period": 60,
"Stat": "Sum"
},
"ReturnData": false
},
{
"Id": "request2",
"MetricStat": {
"Metric": {
"Namespace": "MyService",
"MetricName": "ExternalRequests",
"Dimensions": [
{
"Name": "Component",
"Value": "component-1"
}
]
},
"Period": 60,
"Stat": "Sum"
},
"ReturnData": false
}
]
}
Sample Response
{
"MetricDataResults": [
{
"Id": "errorRate",
"Label": "Error Rate",
"StatusCode": "Complete",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
0.1,
0.5,
0.3
]
},
{
"Id": "errorRatePercent",
"Label": "% Error Rate",
"StatusCode": "Complete",
"Timestamps": [
1518868032,
1518867732,
1518867432
],
"Values": [
10,
50,
30
]
}
]
}
See Also
For more information about using this API in one of the language-specific AWS SDKs, see the following: