Say why no container runtime was found #1571

What happens

When no container runtime can be reached, the error is one line:

Could not find a working container runtime strategy

getContainerRuntimeClient tries five strategies in turn. The reason each one failed is logged at debug level and then discarded. The person who hits this, usually on first use or in a new CI environment, has to rerun with DEBUG=testcontainers* to learn whether the socket was missing, permission was denied or DOCKER_HOST pointed somewhere unreachable. Seven issues in this repo quote the message.

Expected

The error says what was tried and why each attempt failed, and points to the supported runtimes page:

Could not find a working container runtime. Tried:
- TestcontainersHostStrategy: tc.host is not set
- ConfigurationStrategy: DOCKER_HOST is not set
- UnixSocketStrategy: connect ENOENT /var/run/docker.sock
- RootlessUnixSocketStrategy: no socket at /run/user/1000/docker.sock
- NpipeSocketStrategy: not Windows
See https://node.testcontainers.org/supported-container-runtimes/

Where

getContainerRuntimeClient in packages/testcontainers/src/container-runtime/clients/client.ts, and the strategies in packages/testcontainers/src/container-runtime/strategies/.

Notes

• Strategies currently signal "not applicable" by returning nothing, so each needs to return a reason.
• The underlying errors belong in cause.

Dependencies

• Nothing blocks this. It can start now.
• Coordinate with Throw typed errors with a cause #1532: for the error class and cause.