fix: retry transient failures in container image metadata fetch

[Copilot]

Registry requests for container image metadata intermittently fail with timeouts or transient network errors, with no retry logic causing the entire fetch to fail immediately.

Backend (docker-registry.js)

• Added withRetry helper: retries up to 2 times with linear backoff (2 s, 4 s) on transient errors — timeout, ECONNRESET, ECONNREFUSED, ETIMEDOUT, ENOTFOUND, HTTP 5xx
• Non-transient errors (404, auth failures) are rethrown immediately without retrying
• getImageConfig (the entry point for the /metadata API) is wrapped with withRetry

async function withRetry(fn, maxRetries = 2, retryDelay = 2000) {
  for (let attempt = 0; attempt <= maxRetries; attempt++) {
    try {
      return await fn();
    } catch (err) {
      if (attempt === maxRetries) throw err;
      const isTransient = err.message.includes('timeout') || /HTTP 5\d\d/.test(err.message) || ...;
      if (!isTransient) throw err;
      await new Promise(resolve => setTimeout(resolve, retryDelay * (attempt + 1)));
    }
  }
}

async function getImageConfig(registry, repo, tag) {
  return withRetry(() => _getImageConfig(registry, repo, tag));
}

Frontend (ContainerFormPage.tsx)

• When all retries are exhausted, the error message now appends — Use "Fetch metadata" to try again. to direct the user to the existing retry button.

[cmyers-mieweb]

Looks good, let's get this merged