request improperly identified as request to unmonitored host

[rastislav_danis]

we identified issue where some parts of purepath requests are identified as requests to unmonitored hosts.

this environment contains several security zones and monitored hosts are identified by different ip addresses in each zone, so it can happen that call from one zone to another is identified improperly as request to unmonitored hosts.

are there any hints how to handle this setup ?

[ChadTurner]

just to confirm, the oneagent is on the hosts that are coming up as unmonitored... correct? are you still having this issue? If not what did you do to correct this?

[rastislav_danis]

Honestly I can't even remember why I asked that question. It has been covered by dust;)

[ChadTurner]

hahaha understandable, I just wanted to make sure that you dont have the issue anymore and if you did we could work through it. thanks for the update! stay safe!

[istvan_safranek]

I have a familiar issue. There is onagent on the host where two requests are identified as Requests to unmonitored hosts. All the other 100+ requests are fine on this host.

More over these two specific requests are monitored as usually in most cases.

[Julius_Loman · Accepted solution]

This may happen due to adaptive traffic reduction on the agents. If dynatrace does not have the trace (pure path) from the called service, it will be represented as call to unmonitored hosts unless there are some service detection rules defined (or a custom device).

[istvan_safranek]

Hello, thanks for the answer.

On that host there is an apache webserver running. Requests are sent to a weblogic osb cluster (2 managed servers). A multidimensional analysis show me there are 271 individual requests goes to the osb cluster. It is strange to me that only 2 requests are marked as requests to unmonitored requests.

[waikeat_chan]

A very small possibility: But could it be somehow that 2 specific request has went through some load-balancer/network-appliances in-between, before reach the destination-tier? (meaning all other request directly reach destination-tier)

If yes, then maybe somehow those network-appliances has stripped off some Dynatrace-header required for PurePath stitching.

[istvan_safranek]

No, all requests goes through the same loadbalancer and web servers.