Hello, is it possible to deploy DCRUM in a virtual environment? Prospect has his application hosted on Amazon EC2 and they want to monitor it with DCRUM, can the AMD and CAS be installed on virtual machines within the EC2 instance?
If they also have a physical data-center, can an AMD be installed there and the data sent over to the CAS server in the cloud?
While technically possible under some conditions, it might not be the best solution to run in a cloud instance. All depending on how that instance look like. The issue at hand is how you make sure that the AMD get all the network traffic. Slightly related is whether they can run RedHat, but there is usually some dialect of RedHat that work.
Also related to that is the actual network packet capture. Depending on the cloud setup, you might have control of the cloud network carved out to you and you can then do a SPAN/MIRROR or something else on the cloud switch. You also have tthe options to use virtual TAPs (extra cost) if the cloud provider lets you. Realise though that this will effectively double your network traffic budget (more $$$)
Your other question is more straightforward - Yes! You can run the CAS in the cloud and have a AMD in a DC reporting up to the cloud.
What you need to consider though is that the AMD usually has it tentacles in sensitive areas in the DC and you are pushing that raw analytics data out to the cloud - albeit under SSL encryption. It also has implications since the "pump" triggers every 5 minutes and some of the cloud agreements are charged by how much CPU+disk you use as well as how much traffic goes to your instance.
Thanks for the information on DCRUM and Dynatrace components in the cloud, I will work with the prospect following these guidelines.
Apparently British Telecom has implemented a solution with CAS in the cloud already.
I was leading this effort for a CAS in the BT Compute space and did it successfully. Essentially, I just matched the spec and made sure the network had the appropriate access.
The AMDs I haven't had a chance to put in that same space, but I will in the near future. I would follow Kris Z's advice and use a CentOS AMI and then overlay our software over it for the quickest deployment.
I'd be curious to see what machine spec you are looking at using in EC2? Hopefully not a t2.micro ;).
Just a quick note to everyone interested: DC RUM's AMD (and the CAS) can work in the cloud, including Amazon. The only challenge to monitoring cloud-based apps with DC RUM at their hosting location (in the cloud) is to get a copy of the app traffic to to the AMD. There are several ways of doing it.
The simplest, complete, and best documented way it to use virtual tapping solution from established vendors - we recommend Ixia CloudLens (https://www.ixiacom.com/solutions/ixia-cloud-solution).
If your apps have firewalls or load balancers fronting them (usually they do) - check whether this node (e.g. web application firewall - WAF) can feed a copy of traffic it receives to an external receiver. Typically they can, via a GRE tunnel. This tunnel can be terminated on AMD and this way a copy of the app traffic is delivered to the AMD for analysis.
We have Dynatrace DXS consultants who implemented the second solution for customers.
I assume your O 365 traffic does not cross a central point on your LAN or DC edge, like a proxy. In such case, answer depends on where is the O365 hosted. If it's at your provider (like e.g. BT), you have an open path to provision an AMD at the provider. If it's hosted at Microsoft, you won't have a way to provisos an AMD there, or a virtual tap, and you won't have a chance to get the SSL key anyway, so DC RUM's monitoring will be limited in therms of visibility into specific transactions.
My advice would be to use a combination of DC RUM for the LAN users, to see how much are they tasking the LAN with O365 traffic and of what kind, and synthetic monitoring to track Microsoft's availability and performance from the LAN and selected Internet locations. That's the most sensible and economically justifiable solution.
If you are looking for detailed troubleshooting of specific users' performance with the web browser part of the O365, you may think of using a browser agent (former AJAX Edition) to get insight into this specific part of O365 from the browser perspective.
Meenakshi: Hi Kris, As of now I am not aware where Office 365 is hosted. For synthetic monitoring i would like to see the performance test from inside the LAN , do i need to install Private last mile inside customer premises? Also if i record the script with my exchange credentials instead of dummy user . what ensures the security of my exchange credentials.
KZ: Indeed the PLM would be the the most sensible since then you will have same script source and reporting interface for tests ran from both LAN and the Internet.
Regarding the data privacy - your scripts are kept on the Dynatrace Portal and have the credentials encrypted. Agent would decrypt them when executing the script. Synthetic Monitoring specialists would know more in this regard.
No, there is no official documentation as of now. However, you can treat AWS platform as any machine wth OS preinstalled that has been provided to you for installing the AMD. Internally we used the AWS CentOS base image and successfully installed regular AMD software on it for training purposes and CloudLens testing. So it looks that you may take the same path for starter.