#105 output per-database perfdata for pgbouncer pool checks (was |time=0.01 time=0.01 time=0.01 ...)

[georgehansper]

This pull request addresses issue #105

It has the same objectives as the existing pull requests #104 and #76
Sample output:

before:

> sudo -u pgbouncer check_postgres.pl -H /var/run/pgbouncer/ -p 6432 --action=pgb_pool_cl_active -u pgbouncer
POSTGRES_PGB_POOL_CL_ACTIVE OK: DB "pgbouncer" (host:/var/run/pgbouncer/) (port=6432) pgbouncer=1 * test_db1=2 * test_db2=1 | time=0.01s time=0.01s time=0.01s

after:

> sudo -u pgbouncer check_postgres.pl -H /var/run/pgbouncer/ -p 6432 --action=pgb_pool_cl_active -u pgbouncer
POSTGRES_PGB_POOL_CL_ACTIVE OK: DB "pgbouncer" (host:/var/run/pgbouncer/) (port=6432) pgbouncer=1 * test_db1=2 * test_db2=1 | pgbouncer=1; test_db1=2; test_db2=1;

Note that we lose the time=... performance statistic, but this represents the execution time for the query itself, and does not contribute to the actual performance being investigated.

When pgb_pool_maxwait is used, the maxwait time is provided per-database.
This is unrelated to the old time=... performance output, which was the same value repeated over the number of databases.

> sudo -u pgbouncer check_postgres.pl -H /var/run/pgbouncer/ -p 6432 --action=pgb_pool_maxwait -u pgbouncer
POSTGRES_PGB_POOL_MAXWAIT OK: DB "pgbouncer" (host:/var/run/pgbouncer/) (port=6432) pgbouncer=0 * test_db1=0 * test_db2=0 | pgbouncer=0; test_db1=0; test_db2=0;

[machack666]

Thanks for the contribution!