Stabilize enterprise Postgres e2e admin API verification on GKE

[Copilot]

The run-e2e-tests / gke (1.35.5, TestDeployAllInOneEnterprisePostgres$) job was failing after deployment completed successfully, during enterprise verification against the exposed Admin API. The failure came from a single POST /workspaces call timing out behind the GKE load balancer.

• Make enterprise workspace verification retryable

  • Replace the one-shot workspace creation request in verifyEnterpriseWithPostgres() with an Eventually loop.
  • Recreate the HTTP request on each attempt so retries are independent and use the existing Admin API readiness window.

• Avoid false negatives from partially successful attempts

  • Generate a unique workspace name per test run.
  • Accept both 201 Created and 409 Conflict as success, so a prior successful create does not cause a later retry to fail.

• Isolate request behavior behind a helper

  • Extract workspace creation into a small helper that returns:
    • whether the operation should be considered successful
    • the response payload for failure reporting

• Add focused coverage

  • Add helper-level tests for:
    • 201 Created
    • 409 Conflict
    • non-success responses that should continue retrying

require.Eventually(t, func() bool {
    success, response, err := createEnterpriseWorkspace(
        ctx,
        helpers.DefaultHTTPClient(),
        adminURL,
        adminPassword,
        workspaceName,
    )
    if err != nil {
        lastResponse = err.Error()
        return false
    }
    lastResponse = response
    return success
}, adminAPIWait, time.Second)

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}