On this page:
Don't extract the numbers in this guide to sizing for log analytics because the Analytics Agent must be installed on the local machine to capture log analytics.
Based on our testing, the volume of events being sent to the analytics agent is the limiting factor in determining how many APM agents can report to one remote analytics agent.
The tests were conducted on virtual hardware and programmatically generated work load. Real world work loads may vary. To best estimate your hardware sizing requirements, carefully consider the traffic patterns in your application and test in a test environment that closely resembles your production application and user activity.
One business transaction can traverse many tiers. In each tier, one business transaction traverses one node. One node produces one request per business transaction when the transaction is synchronous. For async transactions, multiple events may be generated by a node for a single request. One request equals one analytics event. To calculate how many events a business transaction generates, you need to count the number of tiers/nodes that are sending data into the Analytics Agent.
You can estimate the number of events using the following formula:
One business transaction generates events at a rate = calls per minute times the number of tiers reporting analytics data for the business transaction.
In simple terms: #events for one business transaction = calls per minute times # of tiers.
For complete information on Amazon EC2 instance types, see https://aws.amazon.com/ec2/instance-types/.
The testing was performed using c3.large, c3.xlarge, and c4.4xlarge.
|Model||vCPU||Mem (GiB)||SSD Storage (GB)|
|c3.large||2||3.75||2 x 16|
|c3.xlarge||4||7.5||2 x 40|
|Analytics Agent Host Machine||Analytics Agent events/min||Total CPU%||JVM Heap (Mb)|