Diagnosing outage at same time every day

I have a Cisco router, but a Juniper EX4500 virtual chassis distribution (aggregation?) switch (all layer 2, no routing) and a mix of Juniper and another brand below it. Every switch below has a redundant connection to the VC via LACP and 2 links.

At the same time every day, I start getting notices that some switches below the 4500 won't ping. And in fact, they won't. The switches each have 2 IPs in separate vlans (management vlan and data vlan), both won't ping. Another switch sitting next to it with the same vlan membership but it's own LACP link to the 4500 will ping. When a switch goes down, the "whole thing" goes down. No IPs attached to it will ping. Switches range from 1U RJ45 to a Juniper EX8216.

There is no messaging on the router or the switches (other than "NTP server not available" and stuff like that). Logging into the 4500 shows no STP blocked ports, and LACP is still working fine. I have tried sniffing the traffic but it's hard to get it in the right place at the right time because not all switches under the 4500 go down every time.

How would traffic get black holed for an entire interface if the interface still shows up and running?

Any help would be appreciated.