Reduce spine startup time by obtaining settings in bulk

[TheWitness]

Is your feature request related to a problem? Please describe.

To reduce the startup time of spine, get the settings in bulk with a single call.

Describe the solution you'd like

When you run spine with -V 6 verbosity, you can see output like the following:

Total[0.0072] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'path_php_binary'
Total[0.0074] DEBUG: The path_php variable is /usr/bin/php
Total[0.0074] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'availability_method'
Total[0.0076] DEBUG: The availability_method variable is 0
Total[0.0076] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'ping_recovery_count'
Total[0.0077] DEBUG: The ping_recovery_count variable is 0
Total[0.0077] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'ping_failure_count'
Total[0.0079] DEBUG: The ping_failure_count variable is 0
Total[0.0079] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'ping_method'
Total[0.0080] DEBUG: The ping_method variable is 0
Total[0.0080] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'ping_retries'
Total[0.0082] DEBUG: The ping_retries variable is 0
Total[0.0082] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'ping_timeout'
Total[0.0084] DEBUG: The ping_timeout variable is 400
Total[0.0084] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'snmp_retries'
Total[0.0085] DEBUG: The snmp_retries variable is 3
Total[0.0085] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'log_perror'
Total[0.0087] DEBUG: The log_perror variable is 1
Total[0.0087] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'log_pwarn'
Total[0.0089] DEBUG: The log_pwarn variable is 0
Total[0.0089] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'boost_redirect'
Total[0.0090] DEBUG: The boost_redirect variable is 0
Total[0.0090] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'boost_rrd_update_enable'
Total[0.0093] DEBUG: The boost_rrd_update_enable variable is 0
Total[0.0093] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'log_pstats'
Total[0.0095] DEBUG: The log_pstats variable is 0
Total[0.0095] DEBUG: The threads variable is 64
Total[0.0095] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'poller_interval'
Total[0.0096] DEBUG: The polling interval is 60 seconds
Total[0.0097] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'concurrent_processes'
Total[0.0098] DEBUG: The number of concurrent processes is 1
Total[0.0098] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'script_timeout'
Total[0.0100] DEBUG: The script timeout is 25
Total[0.0100] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'selective_device_debug'
Total[0.0102] DEBUG: The selective_device_debug variable is
Total[0.0102] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'spine_log_level'
Total[0.0104] DEBUG: The spine_log_level variable is 0
Total[0.0104] DEVDBG: SQL:SELECT SQL_NO_CACHE value FROM settings WHERE name = 'php_servers'

Describe alternatives you've considered

A few milliseconds to a second or two can be shaved off the polling time by performing the call in bulk.

Additional context

It's something we introduced in cmd.php and script_server.php some time ago. We should have it here too.