Only start APIs / API requesters, which have been properly loaded

Currently it can happen, that z/OS Connect OpenAPI 3.0 servers start APIs or API requesters, which haven't been properly loaded. The following sample is from a cluster of two z/OS Connect servers. A new API was deployed into the shared apps directory of the two servers, but one z/OS Connect server failed to properly load the new API. According to IBM's analysis this happened due to a "shared storage I/O interruption". Since the deployment automation reacts on the successful start message CWWKZ0001I, the problem during the deployment wasn't detected. Later attempts to invoke REST services of the API failed, because the API was not properly loaded.

[11/17/25, 10:46:57:100 CET] 00000180 bm.ws.app.manager.war.internal.WARDeployedAppInfoFactoryImpl I CWWKZ0133I: The xyz application at the /usr/lpp/zosconnect/runtime/shared/apps/xyz.war location is being expanded to the /usr/lpp/zosconnect/runtime/servers/ZOSCONN1/apps/expanded/xyz.war directory.
[11/17/25, 10:46:58:162 CET] 00000180 com.ibm.ws.logging.internal.impl.IncidentImpl                I FFDC1015I: An FFDC Incident has been created: "java.io.EOFException com.ibm.ws.app.manager.war.internal.WARDeployedAppInfoFactoryImpl 163" at ffdc_25.11.17_10.46.58.0.log
[11/17/25, 10:46:58:166 CET] 00000180 bm.ws.app.manager.war.internal.WARDeployedAppInfoFactoryImpl E CWWKZ0131W: Could not expand the application xyz because of the exception null.
[11/17/25, 10:46:58:679 CET] 00000180 com.ibm.zosconnect.engine.impl.OperationManagerImpl          E BAQR1713E: Failed to access API resources for xyz.war due to server configuration.
[11/17/25, 10:46:58:681 CET] 00000180 com.ibm.zosconnect.engine.impl.ZosAssetManagerImpl           E BAQR1715E: Failed to access API resources for xyz.war due to server configuration.
[...]
[11/17/25, 10:46:58:688 CET] 000000d6 com.ibm.ws.app.manager.AppMessageHelper                      A CWWKZ0001I: Application xyz started in 1.588 seconds.
[...]
[11/18/25, 14:48:17:433 CET] 00001326 com.ibm.zosconnect.engine.impl.ProviderInvokeOperationImpl   E BAQR1672E: An error has occurred when trying to use a z/OS Asset. Error: BAQR1662E: Operation with name %2Fapi%2Fxyz%2Foperation1/post is not available.

We would like this behavior to be changed. z/OS Connect should not start APIs or API Requesters, which haven't been properly loaded. Since the deployment of APIs could end up in a lot of different errors and since other operations may run in the same z/OS Connect server in parallel, it is hard for a deployment automation to handle all potential error scenarios. Because of this when the the successful start message CWWKZ0001I appears, the deployment currently is treated as successful. We would like to be able to rely on this. 

Since this hasn't been accepted as a defect by IBM, we have opened this idea.