Slow queries on remote MySQL?

Question

linux

I'm migrating a project to git and I'm faced with the fact that I need to connect to a remote database in order to work with up-to-date information. I set up a bunch of nginx + php7, but when I try to open a page with queries made through mysqli, I get an error like mysqli::__construct(): (HY000/2002), or the page opens very slowly (from 30 seconds or more). The server itself is located in the data center of Moscow, and I'm in the Trans-Baikal Territory, the ping is about 60ms. There is no difference how to connect to the database using a tunnel or a direct connection.
I tried to set up replication from server to server located in the same data center, but there were also delays

my.cnf

[client]
port                            = 3306
socket                          = /var/lib/mysql/mysql.sock

[mysqld_safe]
nice                            = 0
socket                           = /var/lib/mysql/mysql.sock
pid_file                          = /var/lib/mysql/mysql.pid
log_error                       = /var/log/mysql/error.log

[mysqld]
performance_schema
user                              = mysql
port                              = 3306
datadir                          = /var/lib/mysql
socket                           = /var/lib/mysql/mysql.sock
pid_file                          = /var/lib/mysql/mysql.pid
log_error                       = /var/log/mysql/error.log

skip_external_locking         = 1
skip_name_resolve             = 1
event_scheduler                 = 1

max_join_size                     = 1000000
max_sort_length                 = 1024
max_connections                = 900
max_allowed_packet           = 256M
max_connect_errors            = 1000

wait_timeout                       = 30
interactive_timeout             = 30
net_read_timeout                = 30
net_write_timeout               = 30

key_buffer_size                  = 32M
read_buffer_size                 = 4M
sort_buffer_size                  = 4M
read_rnd_buffer_size          = 4M

table_cache                        = 2G
table_open_cache              = 5000

max_heap_table_size         = 512M
tmp_table_size                   = 512M
tmpdir                                = /temp

query_cache_type              = 0
query_cache_size               = 0
query_cache_limit             = 0

innodb_open_files             = 35000
open_files_limit                 = 35000

join_buffer_size                 = 16777216
thread_cache_size             = 8

slow_query_log                 = 1
long_query_time               = 7.000000
log_queries_not_using_indexes   = 0
slow_query_log_file          = /var/log/mysql/long.log

innodb_read_io_threads          = 64
innodb_write_io_threads         = 64

innodb_buffer_pool_size             = 8G
innodb_buffer_pool_instances    = 8
innodb_lock_wait_timeout          = 60
innodb_log_file_size                    = 256M
innodb_log_files_in_group          = 8
innodb_log_buffer_size               = 16M
innodb_file_per_table                 = 1
innodb_flush_log_at_trx_commit  = 2
innodb_flush_method                = O_DIRECT

transaction-isolation                 = READ-COMMITTED
wsrep_on                                   = OFF

Answer 1

Most likely this is due to the fact that the site and the database are on different servers. I remember doing something like this. Separated the database and the site on different servers for better performance. There was no effect, on the contrary, it became even slower. Although maybe I'm a curmudgeon myself)

Answer 2

If there is a lot of data in the sample, then they simply may not reach due to restrictions, see the config
Solution is so-so
Do replication