27 Aug 2019 06:28 AM - last edited on 18 Oct 2022 03:54 AM by MaciejNeumann
we are using a self-hosted Dynatrace to monitor our cloud system (microservice infrastructure).
Our system is still under development. We plan to have several worldwide data/application centers (e.g. Europe, America, and APAC). What we would like to know is the best scenario - how to organize and monitor that.
Should we go with one Dynatrace instance that will gather all data from across the world or it will be too heavy and we have to setup 3 instances (Europe, America, and APAC)?
As additional info, we plan to have a huge network.
Solved! Go to Solution.
The locations of what you're monitoring will not be the determining factor in how you deploy Dynatrace. It is designed to handle such deployments. What you would want to do is think out your use of ActiveGates which can help simplify things and lower your network footprint through compression of the monitoring data.
Most important will simply be ensuring that your cluster is sized appropriately based on the number of monitored hosts and the amount of traffic that will be flowing through your applications. You can scale both vertically and horizontally. There are some docs to help get an idea of what would be suitable but if this is truly a large deployment you should have a proper sizing and architectural planning session performed with Dynatrace Services.
From my side I can admit that having all regions in single dynatrace may help you with managing configuration. If all will be kept in single environment and splitter by management zones only you can always use single DT configuration for whole world. I mean custom services, request attributes, thresholds, custom alerts etc. you can as well compare regions on single dashboard without need to do it manually.
As James said, you will need activegates as proxy for regions to send data to central server.