Is your SQL instance on the same server that you are rebooting? If so, it does take some time reload the database into memory after a reboot. If it is abnormally slow for the same records though, there may be something else going on.

Are you in a multi-tier setup?

My SQL instance is on a different server. We reboot only the app server, and avoid rebooting the SQL sever. Just weird that after even 12 - 18 hours of the reboot, the application is slow.

Is your application server a VM? It does sound a bit strange that it is slow initially and then runs better after a day or so. It would make sense if it was your SQL server doing this, but you stated that the SQL server isn't rebooted here.

I would say the most likely culprits are perhaps anti-virus running on the application server or some other type of scan happening that is initiated by the reboot. Does task manager give any hints as to high CPU or memory usage by chance? Figure out exactly what is slowing it down such as maybe memory swapping, CPU maxing out, disk activity, etc. If this is a VM you may just need to throw more resources at it.

An application like Veeam ONE or PRTG is extremely useful for figuring out where bottlenecks are and giving you the red flags to act on.

The application server is a VM and we have anti-virus running on the server.

We will start running some stats on the VM.

I would be very curious what happens with regards to performance if you restart the application server and disable AV.

There is a great way to measure performance of Sage X3 (for best results, do this without anyone in the system if possible): https://support.na.sage.com/selfservice/viewdocument.do?noCount=true&externalId=75064&sliceId=1&isLoadPublishedVer=&docType=kc&docTypeID=DT_Article&stateId=13312&cmd=displayKC&dialogID=531427&ViewedDocsListHelper=com.kanisa.apps.common.BaseViewedDocsListHelperImpl&openedFromSearchResults=true

I would run this after a reboot where you know it is known to be slow, and then disable say AV and run it again to see if there is an obvious difference in the results.

I would definitely look into something like PRTG (this is free for up to 100 monitors which would be more than enough for this purpose). We use it all the time for monitoring the performance of our systems and being proactive.