ActiveGate Hardware sizing
Hello. Are there any calculator or spreadsheet for evaluating ActiveGate hardware sizing required ?
Imagine I have 1000 servers, how I kow how many ActiveGate servers will be required ? What hardware will be needed for each ActiveGate for cover demand of those ?
Many Thanks
Best Answer
In general there is no calculator for this. In most case, when you have 1000 hosts you will use multiple AG's because they are in different VLAN's. One of the use case for Envirionment ActiveGate is keeping them close to monitored hosts to simplify network rules (for example). You can use as well ActiveGate than as storage for heap dumps, to execute plugins etc. ActiveGate in general if it is not dedicated to synthetic monitoring does not consume much resources. It's only kind of proxy for traffic from agent. Data are compressed and send too server. For 1000 hosts what I would be aware of is scaling Dynatrace cluster itself (if you have managed, for saas you don't have to think about it).
Sebastian
The information I can not find and am particularly interested in are the sizing guidelines for Cluster ActiveGates that are to handle (external) Agentless RUM, for a large scale deployment (10k+ users daily). Both AG sizing and network requirements.
Anyone can add some information about this?
I ended up here with a similar question. I don't feel like the question has been fully answered here. In the documentation we can find ActiveGate hardware and system requirements with a Sizing guide https://www.dynatrace.com/support/help/setup-and-configuration/dynatrace-activegate/installation/activegate-hardware-and-system-requirements/#sizing-guide. You can expand horizontally (add exta AGs) and/or vertically (add extra hw resources).
The information I can not find and am particularly interested in are the sizing guidelines for Cluster ActiveGates that are to handle (external) Agentless RUM, for a large scale deployment (10k+ users daily). Both AG sizing and network requirements.
@Radoslaw S. perhaps you can add some information about this?
What I can share only that it's really hard to estimate that as there are multiple variables there. What is worth to share is that CPU of Cluster AG < 60%. We load test Cluster AG with 4k user actions / sec on 2x AWS c5.2xlarge
Hope this helps.
