APM Server Fleet managed vs Legacy: incompatible?

[Attila_Szeremi]

In our company with APM before Fleet, we used to not only view transactions in APM, but also create dashboard charts about the transactions in Kibana; statistics about how fast HTTP requests take and things like that.
For this we would make sure with Lifecycle Policies that transactions never get deleted, but spans would get deleted if they were older than 30 days, to save space.

Now with Fleet, I'm seeing that there are now this thing called traces where the transactions and spans are both in the same index?

So it looks like I can't do the same thing I did before.
I can't keep just the transactions forever and discard spans after 30 days, because they're not in separate indexes anymore. But the spans take up a lot of space, so I couldn't just keep them either.
Also, I'm worried that I can't make the same kind of Kibana charts as I used to (or at least not have them be as fast as they used to), because all the spans are also part of the same index (and we don't care about spans when showing charts), and trying to make aggregating queries runtime on these traces would be really slow, because it has a lot of spans in them.

So is it really the case that APM Server with Fleet works completely differently, and they're not compatible in the sense that I can't create charts with the new data?

[simitt (Silvia Mitter)]

Hi @Attila_Szeremi ,

from 8.0 onwards, traces and spans are indeed stored in the same traces-apm* datastream, independent of whether apm-server is run as binary on its own or managed vie Elastic Agent & Fleet. Alongside this change, aggregated metrics were introduced. The metrics are powering large portions of the APM UI and have a longer retention period by default. The metrics are calculated from trace events, along various dimensions and might be what you are looking for to keep a rolled up view of how the services were behaving over time. Let us know if this works for you.