Continue check execution when only a few vhosts are unhealthy

[FlorianVeaux]

Currently the check for aliveness-test was failing the whole check whenever a single vhost was reporting unhealthy (a non 2xx status code).

This PR fixes that behavior by gracefully handling the error, marking that vhost unhealthy but continuing the rest of the execution.

[codecov]

Codecov Report

Merging #6954 into master will increase coverage by 4.24%.
The diff coverage is 100.00%.

Impacted Files	Coverage Δ
rabbitmq/datadog_checks/rabbitmq/rabbitmq.py	84.61% <100.00%> (ø)
rabbitmq/tests/test_unit.py	97.27% <100.00%> (ø)
mapr/tests/test_unit.py
php_fpm/datadog_checks/php_fpm/__init__.py
datadog_checks_base/tests/test_warnings_util.py
hivemq/tests/conftest.py
aerospike/datadog_checks/aerospike/__about__.py
consul/tests/test_unit.py
datadog_checks_base/tests/test_proxy.py
...yototycoon/datadog_checks/kyototycoon/__about__.py
... and 852 more