Make container port mapping timeout configurable

[khalilswdp]

During container startup, GenericContainer polls inspectContainer until the Docker
daemon reports all exposed ports as mapped, with a hard-coded 5-second limit. On heavily
loaded hosts (shared CI runners in particular) the daemon can take longer, and the
resulting failure is an opaque Awaitility timeout rather than the detailed diagnostics
produced for later startup failures.

Changes

• Add a container.port.mapping.timeout configuration property (default: 5 seconds, so
  out-of-the-box behavior is unchanged), read via the standard mechanism — env var
  TESTCONTAINERS_CONTAINER_PORT_MAPPING_TIMEOUT, ~/.testcontainers.properties, or a
  classpath testcontainers.properties.
• Use it in GenericContainer.tryStart() in place of the hard-coded value.
• Document the property in docs/features/configuration.md and cover the
  default/property/env-var precedence in TestcontainersConfigurationTest.

A configuration property (rather than a per-container API) was chosen deliberately: the
latency this timeout absorbs is a characteristic of the Docker daemon and host load, not
of any particular container, mirroring existing environment-level knobs such as
client.ping.timeout.

[eddumelendez]

Closing. See #11938 (comment)