chore(ci): wait for server in integration test runs after config restore (#3540)

The integration workflow can start policy-service BATS while the
background OpenTDF server is still restarting after tdf-roundtrips
restores opentdf.yaml. That leaves the next step racing a brief
connection-refused window.

This waits for the restored config to be picked up and the server health
check to return before tdf-roundtrips exits, so the following BATS step
starts against a ready server.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Tests**
* Improved test suite stability by ensuring services are fully ready
after restore: added a brief initialization pause, active health polling
before continuing, and explicit failure if the service never becomes
healthy.

<!-- review_stack_entry_start -->

[![Review Change
Stack](https://storage.googleapis.com/coderabbit_public_assets/review-stack-in-coderabbit-ui.svg)](https://app.coderabbit.ai/change-stack/opentdf/platform/pull/3540?utm_source=github_walkthrough&utm_medium=github&utm_campaign=change_stack)

<!-- review_stack_entry_end -->
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: jakedoublev <jake.vanvorhis@virtru.com>

Author: jakedoublev

test/tdf-roundtrips.bats

@@ -159,6 +159,8 @@ wait_for_green() {
     fi
     sleep 4
   done
+  echo "WARNING: service failed health check after sleep and polling" >&2
+  return 1
 }
 
 write_opentdf_config() {
@@ -357,5 +359,7 @@ teardown() {
 teardown_file() {
   if [ -f opentdf-test-backup.yaml.bak ]; then
     mv opentdf-test-backup.yaml.bak opentdf.yaml
+    sleep 4
+    wait_for_green
   fi
 }