We have a client that was on FreePBXv13. They used UCP extensively; most significantly the manager of the health care center uses UCP to view call history and voicemail by extension.
We migrated them from v13 to v15. UCP in v15 is very slow and uses LOTS of CPU cycles. We host in AWS. Had FreePBXv13 running for a long time, never an issue, running AWS t2.micro - never an issue even once. Now on V15 (backup v13 / restore v15) we have constant CPU cycle issues, due to UCP usage. I “bought up” the machine type several “notches” to a t2.medium; we run a 500 extension v13 on medium with room to spare, this medical clinic is just 20 users, a t2.meduim should be overkill by an order of magnitude. The PBX was sitting at a load factor of .2 this morning. The manager came in and logged into UCP, soon their voice quality was poor. They called into to complain of bad voice quality and slow UCP. The load factor on the PBX was over 3 for 1 min, 5 min and 15 min. I removed UCP as an option for them. They are not happy with me. As soon as I stopped all UCP and removed their access, the PBX load went down to .3.
CDRs are all indexed.
If we use the free asternic CDR report utility, CDR come up very very fast (http://www.asternic.biz/asternic-cdr-reports.php - sorry if I violated some standard by placing a URL in the post!!). Based on the fast asternic CDR I am not inclined to think the issue is the CDR database. Also we trim the DB to only be the last 6 months of CDR.
Do not know what to do and do not want to loose a client over poor UCP app that makes the whole PBX bad.