How to run a metric query on a devicegroup containing all devices?

Hello, a customer is attempting to run a metric query on a device group containing all devices.
It is a large environment and would be polling ~22 EDAs. The query is being run on the ECA. The query would primarily be on the bytes_in and bytes_out metric in the net_detail category and would use the /metrics/totalbyobject endpoint for the past 24h.
When attempting to run the query, the query returns results, but stops with Results Incomplete at about 6 minutes.
Is there a limit on how long the query can run? Is there a way to increase that limit if there is one?
Are there any performance or stability concerns on doing such a query on such a large device group?
Is there a better way to approach this use case?
Thank you in advance.

Have you tried using stats from the network object, or from the “All Activity” application?

I have not. I will recommend this to the customer.
To expand more on the use-case, the customer is looking to search devices by their peers. Meaning, they would like to take a peer IP(not a discovered device) and see which discovered devices are talking to it.
Records would not be an option as records do not cover all of the traffic in the environment.
Thank you

They may need to query the metrics in chunks and build a local index for that use-case.