Postgresql DB connection through Haproxy

[Stayler]

I want to use Haproxy as load-balancer for Zabbix-DB connection:
I setup haproxy & custom check.
And I do not have problems with checks, my node is always UP, but in time to time(once in 1-2 days, in random time) a get message, "Zabbix database is not available.", in logs:

 16708:20200224:202412.152 [Z3001] connection to database 'zabbix' failed: [0] server closed the connection unexpectedly
        This probably means the server terminated abnormally
        before or while processing the request.

Can you help me, where to look?

#> haproxy -v
HA-Proxy version 1.8.17 2019/01/08

Haproxy.conf:

global
  log 127.0.0.1:514 local3
  pidfile /run/haproxy.pid

  maxconn 30000
  nbproc  6
  cpu-map 1 0
  cpu-map 2 1
  cpu-map 3 2
  cpu-map 4 3
  cpu-map 5 4
  cpu-map 6 5
  stats bind-process 6
  stats socket /run/haproxy/admin.sock mode 660 level admin
  user  haproxy
  group haproxy
  external-check
  daemon


defaults
  mode  http
  log   global
  option  dontlognull
  retries 3
  maxconn 30000

#
# Endpoints
#

#PostgreSql-Write

listen PostgreSql-Write
  bind *:19307
  mode tcp
  balance leastconn
  timeout client  24h
  timeout server  24h
  timeout connect  3s
  log-format '{"hostname":"%H","type":"haproxy","date_time":"%t","timestamp":"%Ts","format":"tcp","remote_addr":"%ci","frontend_port":"%fp","backend_name":"%b","upstream_addr":"%si","bytes_read":"%B","bytes_uploaded":"%U","retries":"%rc","upstream_connect_time":"%Tc","session_duration":"%Tt","termination_state":"%ts","handshake_time":"%Th"}'
  acl is-PostgreSql-Write-dead nbsrv(PostgreSql-Write) gt 1
  tcp-request content reject if is-PostgreSql-Write-dead
  option external-check
  external-check command /etc/haproxy/zabbix_psql_master.sh
  server DB1 <IP>:<PORT> check inter 3s fall 2 rise 2 on-marked-down shutdown-sessions
  server DB2 <IP>:<PORT> check inter 3s fall 2 rise 2 on-marked-down shutdown-sessions