25 Aug 2021 01:06 PM - last edited on 07 Sep 2021 01:57 AM by MaciejNeumann
What is the maximum number of servers in a monitoring environment?
Currently we have 30000+ servers in 15+ monitoring environment. Cross environment flow will be cumbersome since we have to make 15x15 remote connections. We are planning to merge all environment together. Can we put all the 30000+ servers(not host units) in one monitoring environment?
If we a hard limitation there, Is there any enhancements coming in the product to remove this?
Solved! Go to Solution.
@awahab This is a very very good question. Ultimately, it boils down to platform. I see you are tagged as managed so that good as you hold the keys to the resource and size of your Clustered Hosts. Take a look at the following link as it will help guide you with sizing the Managed Cluster Hosts and the Activegate(s) that you may have spun up: https://www.dynatrace.com/support/help/setup-and-configuration/dynatrace-managed/installation/dynatr...
If you were SaaS then ultimately the sizing issues with be on Dynatrace Corp. to ensure that your environment is running healthy. Both Methods have their pros and cons.
You can always reach out to Dynatrace Sales Reps or even support on the sizing of really large monitoring footprints.
There is no hard limit AFAIK. As long as you have your cluster nodes sized properly - which I believe you already have. Since you have those agents already installed, then it's a kind of a proper migration plan.
Anyway - a press release from 2 years ago mentions the scalability for at least 50k hosts.
Maybe @Radoslaw_Szulgo can provide a more accurate answer.
That's correct. We do not recommend running more than 25k of a typical load in a single environment. Of course it may vary, depends on what's the specific of what you monitor. A single cluster can handle even more than 50k OS agents and we currently support up to 30 cluster nodes.
I think my question is misunderstood. I am not asking how many agents can be in a Dynatrace physical environment. The question is about the monitoring environment. This is in Cluster Management -> Environments Screen.
Note: I am not sure why the answer is accepted as solution without consent from the requester btw 🙂
I understood your question. Considering the Dynatrace architecture, it basically is the limit of how many OneAgents can your Managed cluster itself handle. Anyway - it might be a good idea to get your Dynatrace validated by Dynatrace support.
The cluster currently handles 30000 servers and it can handle more. These 30000 servers are spread across those 17 monitoring environments. The environment which has 6000 servers doenst perform very well in terms of rendering the smartscape topology view or finding the root cause DAVIS AI dependencies or if you create a honey comb view of the host or top 15 bad servers consuming CPU/Memory etc.
I know about this limitation when we sized our cluster and our monitoring environment(s) when we did 2 years back. The question is whether this limitation is removed or will be removed in the future or if anyone solved this using different technique.
Do you use Management zones properly and other filters to filter out only data you need on Dashboards and in other parts of Dynatrace?
Querying a lot of data to display has its price. Having a honeycomb for thousands of entities or displaying a whole smartscape topology is possible, but filtering it on MZ is definitely a must.
These are special use cases which may constitute 10-15%
Data Center Migration - Want to study how the servers interact and what services they interact.
What is the top 15 bad Windows servers in terms of consuming resources for a particular process
What is the top 15 bad Linux servers in terms of consuming resources for a particular process
In my experience with Dynatrace Managed, you need to re-evaluate your architecture design, probably for a single environment your need more Active Gates supporting all the traffic from your OneAgents (monitored servers), also your cluster maybe needs to increase the capacity.
What I recommend you is to talk with our Technical Support Engineers or the Sales Reps Team for helping you with the sizing and the recommendations for a large environment architecture.