Query AWS CloudWatch in Java
2017-07-09
If you run a service on the T2 tier of AWS, your machines have a specific compute budget. So, for example, you may want to delay certain CPU-intensive background tasks to a later time, throttling them by the CPUCreditBalance your machine currently has.
The current (averaged) CPUCreditBalance can be retrieved via CloudWatch, but I found the Java API rather unforgiving1 – it fails silently when certain elements like the unit (StandardUnit.Count) are missing from the API call. Here is a snippet that retrieves the last (average) CPUCreditBalance, as a starting point:
String ec2InstanceId = EC2MetadataUtils.getInstanceId();
LocalDateTime endTime = LocalDateTime.now();
LocalDateTime startTime = endTime.minusMinutes(120);
GetMetricStatisticsRequest cpuCreditBalanceRequest = new GetMetricStatisticsRequest()
.withNamespace("AWS/EC2")
.withStatistics(Statistic.Average)
.withMetricName("CPUCreditBalance")
.withPeriod(60)
.withUnit(StandardUnit.Count)
.withStartTime(Date.from(startTime.toInstant(ZoneOffset.UTC)))
.withEndTime(Date.from(endTime.toInstant(ZoneOffset.UTC)))
.withDimensions(new Dimension().withName("InstanceId").withValue(ec2InstanceId));
GetMetricStatisticsResult cpuCreditBalanceResult = getAmazonCloudWatch().get().getMetricStatistics(cpuCreditBalanceRequest);
List<Datapoint> metricData = cpuCreditBalanceResult.getDatapoints();
if (metricData == null || metricData.isEmpty()) {
LOG.warn("Metric 'CPUCreditBalance' is not available.");
} else {
Datapoint mostRecentData = metricData.get(metricData.size() - 1);
Double result = mostRecentData.getAverage();
LOG.info("Latest average of 'CPUCreditBalance': " + result);
}