Our dynatrace server is hosted in the USA. We are coming up with a new AWS environment which will be hosted in Australia.
The server is not an AWS instance.
We are planning to create 1 collector instance in AWS Australia and connect it to a non-AWS server hosted in the USA. The agents will be in the same network as the collector.
Will this infrastructure result in any latency issues in terms of purepath and metrics capture etc.,? If yes, what will be the overhead?
Note: The same infrastructure is being planned for AWS Germany as well.
The scenario you explained is totally supported by Dynatrace AppMon. The short formula between agents and collector is low latency and between collector and server high bandwidth.
I would recommend you to read the below best practices about Network bandwidth, Collector and Server for better understanding.
In typical scenarios, every 500 tps sent from a Collector to an AppMon Server contributes 8 MBit/sec (1 MByte/sec) of upstream bandwidth. The downstream is negligible in comparison. Using compression reduces this value by a factor of approximately 4, but be aware of increased CPU demand on the Collector and AppMon Server when compression is used.
To measure the bandwidth between the Collector and Server, install the jPerf server on the Collector or on the AppMon Server.
Review the below link for more insight.