I am still a newbe in the using Dynatrace and I am fighting with sort our services in the Transaction Flow Dashbord.
I have hostnames starting
and try to get those servers into a observed tier "Sage".
I configured the following rule in the "Configure observed tier" window from the Transaction Flow Dashboard:
* "Host" "contains" "sage*"
* Service and Instance name is "Sage" and "Sage"
* Vendertype is generic
* enable is activated.
What I see is, unfortunately, that the servers aren't appear in the same tear.
Does anyone have an idea, where I should look for that and fix it?
Solved! Go to Solution.
It sounds like you would like your sage app and sage servers to be displayed in the same tier on the transaction flow dashlet but they are separated, correct?
This could be the result of the servers communicating with one another and resulting in them being split into two tiers. Dynatrace automatically identifies the role of the servers you add to the observed tiers and will place them in the location that it deems best so if your it is separating your sage app and sage servers it is because it most likely thinks your sage servers are fed information from your sage app servers and wanted to visualize it better.
Good Morning, David (well, in Germany do we have morning). ,-)
I think you are right. I went deeper into the clarification, why I do not find the server in my agent group but as external services.
I there a nice way to track down, why Dynatrace does not put agents/servers into a specific agent group?
Thanks in advance and have a nice day.
Good Morning Jan-Hendrik, (Thankfully, we have mornings here as well :))
For the server to be a part of a specific agent group it would need an agent to be deployed on it and that agent must be assigned to the agent group in the system profile settings.
In your case, I am guessing that the instance name entry could also be breaking them into multiple tiers. When I took another look at the documentation (https://community.dynatrace.com/community/display/DOCDT65/Observed+Tiers) the instance name was for further granularity. You can try removing it and seeing whether it places all your sage servers in the same tier.
Beyond that, I am not sure how you could dig further besides opening a support ticket to ask the team.
well, I think, you are right.
The situation here at my company is, that we work on sorting the Dynatrace monitoring. We have many unknown entry points and many external call services, but we are pretty sure, that we know all the application.
I realized, that the moving the external service call into an observed tier is not what I really want, but what you said, that I need agents at these points. We realize, that the instrumentalization does not catch them up and sort them.
How am I able to track these cases down to fix it right?
Thank you for your time. .-)
If you have access to the servers where the external services are hosted then you can install the respective Dynatrace agents and assign them to an agent group in your system profile. Here is a really good explanation of the benefits of agent groups: https://www.dynatrace.com/support/doc/appmon/application-monitoring/agents/agent-groups/
If you don't have access to the servers that run your external services then you will have to use the observed tiers to display them in the transaction flow.
An additional measure that you can take regardless of which option you go with, is to setup application definitions within Dynatrace so that you can have a logical breakdown of all the traffic when you are performing an analysis on the data. Here is some information on applications and their uses: https://www.dynatrace.com/support/doc/appmon/administration/application-environment-administration/system-profile/edit-a-system-profile/system-profile-applications/
thank you, the second links is interesting and I will dig into it.
What I found out here is, that we have loadbalancers handling the access to some services. This yields to the situation, that I install and configure the agents (in this case Apache an PHP) and what I see is, that the one transaction goes to the service name (configured on the loadbalancer) and one transaction goes to the php application on the final server.
I am sure, that both transaction are the same (or from the same call), but I do not have the possibility to install one agents on our loadbalance (in this case a keepalived).
Do anyone have an idea to handle this situation?
Beside that do we have some calls going out to service providers or partners and I am fine, if Dynatrace sort them into external calls. .-)
That is the situation where the observed tiers shine. They allow you to designate a tier that your traffic passes through, that cannot be instrumented. You can add the load balancer as a tier and this should allow your transaction flow to look more like reality.
Yes, you actually want the calls to your service providers and partners to be marked as external calls because that is the correct designation for them.
I have to say thank you for your time. .-)
We played around with the observed tiers and it works. *yeah*
So, I think, I had some learnings from this discussion and I am able to sort the transactions and servers in a first step. The sensors will follow afterwards, if we have the big picture complete.
Thank you very much for your help.
Best regards and a nice day.