HA Cluster issue on Secondary FW - States growing and crashing FW2

[uhscale]

Hello

we have running HA Cluster, and found curious issue:

FW1 is running fine with States around 2% of 25 000 000 - in use round about 500 000.

FW2 has running States with 5% of 25 000 000 and is growing slowly to maximum and then FW2 is crashing.
After reboot of FW2 its running fine for a while.

Why is States sync not working on FW2 ?

Has someone similar issue?

Config:
2.4.5-RELEASE-p1 (amd64)
Intel(R) Xeon(R) Silver 4214 CPU @ 2.20GHz
24 CPUs: 1 package(s) x 12 core(s) x 2 hardware threads
AES-NI CPU Crypto: Yes (active)
32GB Ram

Thanks