During bosh create-env, the health monitor may start before nats_sync has
propagated the HM credentials to NATS. This causes NATS connection failures
(AuthError/ConnectError) that result in the health monitor crashing and
being restarted by BPM/Monit.

This change adds retry logic using Bosh::Common.retryable to wait for NATS
to become available, similar to the database connection retry logic added
for the director.

Changes:
- Add bosh-common dependency to bosh-monitor gem
- Add wait_for_nats_connection method with configurable timeout
- Add hm.nats.connection_wait_timeout property (default: 60s)
- Update health_monitor package to include bosh-common gem
- Add unit tests for NATS connection retry behavior

During bosh create-env, nats-sync may start before the director API is
available, causing connection refused errors. This adds retry logic
using Bosh::Common.retryable to wait for the director to become
available before attempting to sync users.

Changes:
- Add bosh-common dependency to bosh-nats-sync for retryable
- Add wait_for_director_connection method with configurable timeout
- Add nats-sync.director.connection_wait_timeout property (default: 60s)
- Handle network errors: ECONNREFUSED, ECONNRESET, ETIMEDOUT, etc.
- Update nats package to build all gemspecs like health_monitor does
- Add unit tests for director connection retry scenarios

During bosh create-env, the scheduler process may start before the
director has completed running migrations, causing 'relation does not
exist' errors when the scheduler tries to query the director_attributes
table.

This adds migration waiting to the scheduler binary, following the same
pattern used by sync_dns_scheduler and metrics_collector. The scheduler
now:
1. Loads only config initially (not full director with models)
2. Calls DBMigrator.ensure_migrated! to wait for migrations
3. Then loads full director and starts the scheduler

Also adds scheduler_logger method to Config for consistent logging.

Update template spec fixtures to include the new connection_wait_timeout
properties that were added for startup race condition resilience:
- director.db.connection_wait_timeout in director.yml.erb_spec.rb
- hm.nats.connection_wait_timeout in health_monitor_templates_spec.rb
- nats-sync.director.connection_wait_timeout in nats_templates_spec.rb