mariadb 10.1 crashing

Hello,
recently I switched to 10.1 and centos7
however, I had to urgently downgrade to 10.0

I faced few problems:

1. master crashing
my master servers are mariadb 10.1 with blackhole tables and around 20k qps
under load spikes (from many web servers, like if web app was restarted) mariadb 10.1 crashes. It does not matter if load is 20k qps or 5k qps.
More than that, it seems related to number of IPs (web app servers) that work with master since a different server that is still not downgraded (because it's production and never crashed) has fewer web app IPs that access it. Few other masters crashed from time to time without any information in logs. systemd notices mariadb is down and restarts it.
After few tries/changes I decided to downgrade to 10.0 and it's stable for a few days already no matter what happens on web app servers (the single 10.1 master is still alive but I am afraid it may begin crashing when I'll add more web app servers)
All masters are the same.

2. Backup slaves crashing
I have few backup slaves that replicate from those masters to myisam with disabled indexes. Single slave replicates from 3 masters, 2 production (i've described above) and 1 different master that's using innodb. I am using parallel replication. Out of sudden (I have few slaves like that) different slave crashing and restarts. I can't tell for sure if this happens as result from backups that run daily, but it's occasional and I've already downgraded it to 10.0

I seen nothing in dmesg, logs, or anywhere else. I know I need to compile/install some debug version and tune my.cnf
I'd like to help you debug this and need some guidance. I am ready to let production server crash once again.

But for now all systems run with 10.0 without any problem or any crash seen so far (maybe this will happen some day and it relates to centos7 or other settings I've missed..who knows)

Thanks
Alex