question

Kalle L. avatar image
Kalle L. asked ·

ESM Proxy Agent setup

Hi,

Couldn't find any earlier discussion about this. When registering a new agent from the agent's side (agent-initiated connection), a proxy agent can be defined:

https://community.dynatrace.com/community/pages/vi...

Is there any further documentation regarding this? What does the proxy agent setup entail, is it basically the same as installing a regular agent but with the exception that this one also accepts and forwards measurement data from one or more agents & sends it to the agent manager? Is it then dedicated to be used as a proxy agent, or can it still execute its own monitoring and forward all the data (own + other agents') to the manager? And finally, is the process of adding the proxy agent as a new agent in the Synthetic Console the same as if it was a regular agent?

One way to achieve similar functionality as the agent proxy would be to define a child agent manager, but I'm wondering whether this type of chain is possible:

Manager A = Parent

Manager B = Child to Manager A

Manager C = Child to Manager B?

So then C would send monitoring data to B, and B would forward it to A, and from A it would go to CAS. Is that type of setup possible?

Br,

Kalle

esm
2 comments
10 |2000000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 50.0 MiB each and 250.0 MiB total.

Oh and about the proxy settings during registration:

The connection is probably http://hostname + port 9016 or https://hostname + port 9017

What about the user ID + password definitions, seeing as the way the agents to agent manager data transfer works is that no login credentials are required - why is this different? So the process of data handling (incoming + outgoing) is probably different then for a proxy, compared to e.g. a child agent manager.

0 Likes 0 · ·

HI anonymous user,

Replying to your main question:

Synthetic Agent & Agent Manager are different components with their separate tasks and they have a different executable to run (while installing).

Agents are not able to receive data from any other agent and they cannot act like a child Agent Manager.

So according to my understanding , below setup as mentioned by you is not possible:

Manager A = Parent

Manager B = Child to Manager A

Manager C = Child to Manager B?

Also regarding the Proxy details in adding an agent manager (wile setting up an agent initiated communication) : it is just a matter of communication from Agent to Agent manager.

If direct communication is not possible but can be established via proxy , then you need to use proxy.

It is exactly the same concept when you do the reverse: agent manager initiated communication . If agent manager cannot communicate directly to agent , then you need to use proxy to achieve the same

Thanks

Himanshu Mor

1 Like 1 · ·
Carol O. avatar image
Carol O. answered ·

Hi. I have modified the documentation for Adding an Agent with Agent-Initiated Communication; I think it was confusing. I'm not sure what you meant by "proxy agent", but the Agent-initiated registration really refers to defining a proxy that may be required for the Agent to communicate with the Agent Manager. You can use this if you need a proxy. It is similar to configuring a proxy server in the Agent General Tab or in the Agent Manager General Tab. You may also want to refer to the proxy server documentation in Servers Content Pane

I think some customers have used multiple hierarchical Agent Managers. If needed, you should be able to use this type of chain (PAM = Parent Agent Manager; CAM = Child Agent Manager):

PAM A = Parent AM with configuration

PAM B = Parent AM reporting to its Parent AM A

CAM C = Child AM reporting to its Parent AM B.

So then C's Agents would send monitoring data to C; C would send data to B, and B would forward it to A.

1 comment Share
10 |2000000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 50.0 MiB each and 250.0 MiB total.

Thank you for the responses Carol and Himanshu! This really cleared it up for me, I know now how to proceed with planning our setup.

0 Likes 0 · ·
Kalle L. avatar image
Kalle L. answered ·

No responses yet, but in the meantime another question: the agents connecting to the proxy need to get their Recorder licensing from somewhere. I suppose you could install DLM on a Proxy Agent and have the agents get their licensing from there, right? So far we've had the DLM running either at CAS or an Agent Manager, but I suppose there's no limitation to put it on a Proxy Agent...

2 comments Share
10 |2000000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 50.0 MiB each and 250.0 MiB total.

HI anonymous user , Response to your last reply about the local license installation using DLM on the proxy agent: yes that works fine as I have already done this in my environment.

Only concern is that if you use a local license file on the agent machine , you will have ot upgrade the license file manually on all such agents who are not fetching the license from central license server(either CAS or agent manager , depends where you have installed licensing server).

1 Like 1 · ·

Hi, please note that you only need the Recorder license when you are using the Recorder GUI. A run-time license is not required for the Recorder; so an Agent does not necessarily need a license. Thanks!

0 Likes 0 · ·