AMS to AEMaaCS Migration - Performing certain job on Publish tier

hi ​@helloosuman,

if possible I’d leverage the author instance as single orchestrator for sending content to the external application because it would really simplify the architecture logic and possible corner cases.

@Component(
    service = EventHandler.class,
    immediate = true,
    property = {
        EventConstants.EVENT_TOPIC + "=org/apache/sling/distribution/agent/package/distributed"
    }
)
public class TopicReplicationHandler implements EventHandler {

    @Override
    public void handleEvent(Event event) {
        // Extract path and verify it's a Topic
        String[] paths = (String[]) event.getProperty("paths");

        // Verify it's an activation/publish event
        // Send to external application from Author (only once)
        sendToExternalApp(paths);
    }

    private void sendToExternalApp(String[] paths) {
        // Your external API call logic here
        // This executes only once from Author
    }
}

If Author is not viable then you can leverage Discovery Service to find the topology leader within a EventHandler:

@Component(
    service = EventHandler.class,
    immediate = true,
    property = {
    	EventConstants.EVENT_TOPIC + "=" + ReplicationAction.EVENT_TOPIC
    }
)
public class LeaderElectedTopicHandler implements EventHandler {

    private static final Logger LOG = LoggerFactory.getLogger(LeaderElectedTopicHandler.class);

    @Reference private DiscoveryService discoveryService;

    @Override
    public void handleEvent(Event event) {
        // Only process if this instance is the leader
        if (!isLeaderInstance()) {
            LOG.debug("Skipping event - not the leader");
            return;
        }

        // Process replication event and send to external app
        sendToExternalApp(event);
    }

     /**
     * Check if this instance is the cluster leader
     */
    private boolean isLeaderInstance() {
        try {
            TopologyView topology = discoveryService.getTopology();
            if (topology == null) {
                LOG.warn("Topology is null - cannot determine leader status");
                return false;
            }

            InstanceDescription localInstance = topology.getLocalInstance();
            if (localInstance == null) {
                LOG.warn("Local instance is null - cannot determine leader status");
                return false;
            }

            boolean isLeader = localInstance.isLeader();
            LOG.debug("Leader status: {}, Sling ID: {}", 
                     isLeader, localInstance.getSlingId());

            return isLeader;

        } catch (Exception e) {
            LOG.error("Error determining leader status", e);
            return false;
        }
    }
}

This should works but, as corner case to evaluate, the topology can change during deployments and there may be brief periods with no leader or multiple leaders during transitions.