/replication-orchestrator

name: replication-orchestrator
description: |
  Orchestrates end-to-end replication workflows spanning multiple concerns: new environment setup,
  production incident response, and performance optimization for AEM 6.5 LTS.
license: Apache-2.0

Replication Orchestrator

Coordinates complex replication workflows that span multiple sub-skills (configure, replicate, troubleshoot).

When to Use This Skill

Use the orchestrator for multi-step scenarios requiring coordination across sub-skills:

• **New Environment Setup:** Configure agents → Test replication → Troubleshoot issues
• **Production Incidents:** Diagnose problem → Isolate root cause → Fix and verify
• **Performance Optimization:** Monitor metrics → Tune configuration → Validate improvements
• **Migration Preparation:** Audit current setup → Document dependencies → Plan cutover

For single-concern tasks, use the specific sub-skill directly instead of the orchestrator.

Workflow 1: New Environment Setup

End-to-end workflow for setting up replication in a new AEM 6.5 LTS environment.

Prerequisites

• Author instance running and accessible
• Publish instance(s) running and accessible
• Dispatcher installed and configured
• Service user accounts created
• Network connectivity verified

Steps

1. Configure Default Replication Agent

**Delegate to:** [configure-replication-agent](../configure-replication-agent/SKILL.md)

**Actions:** 1. Create default replication agent on Author 2. Configure transport URI: `http://publish-host:4503/bin/receive?sling:authRequestLogin=1` 3. Set service user credentials 4. Enable the agent

**Verification Checkpoint:**

# Test agent connectivity
curl -u $AEM_USER:$AEM_PASSWORD \
  http://localhost:4502/etc/replication/agents.author/<agent-name>.test.html

Expected: "Replication test succeeded"

2. Configure Dispatcher Flush Agent

**Delegate to:** [configure-replication-agent](../configure-replication-agent/SKILL.md)

**Actions:** 1. Create flush agent on each Publish instance 2. Configure transport URI: `http://dispatcher-host:80/dispatcher/invalidate.cache` 3. Set serialization type to "Dispatcher Flush" 4. Enable the agent

**Verification Checkpoint:**

# Test flush agent connectivity
curl -u $AEM_USER:$AEM_PASSWORD \
  http://publish-host:4503/etc/replication/agents.publish/flush.test.html

Expected: "Replication (Dispatcher Flush) test succeeded"

3. Test Content Replication

**Delegate to:** [replicate-content](../replicate-content/SKILL.md)

**Actions:** 1. Create test page: `/content/test/replication-check` 2. Activate via Quick Publish 3. Verify on Publish instance 4. Verify Dispatcher cache invalidation

**Verification Checkpoint:**

# Check page on Publish
curl http://publish-host:4503/content/test/replication-check.html

# Check page on Dispatcher
curl http://dispatcher-host:80/content/test/replication-check.html

# Verify cache was invalidated (should see fresh content)

Expected: Page content identical on all instances

4. Configure Monitoring

**Actions:** 1. Enable JMX monitoring for queue metrics 2. Set up log monitoring for replication errors 3. Configure alerts for queue depth > 20 items 4. Document runbook for common issues

**JMX Bean:**

com.day.cq.replication:type=Agent,id=<agent-name>
  - QueueNumEntries
  - QueueBlocked
  - QueueProcessingSince

5. Handle Any Issues

**If problems occur, delegate to:** [troubleshoot-replication](../troubleshoot-replication/SKILL.md)

**Common setup issues:**

• Connection refused → Verify target instance running and network connectivity
• 401 Unauthorized → Check service user credentials
• Queue blocked → Review error.log for root cause
• Content not appearing → Check Dispatcher cache invalidation

Success Criteria

• [ ] Default replication agent enabled and passing test
• [ ] Dispatcher flush agent enabled and passing test
• [ ] Test page successfully replicated to Publish
• [ ] Test page accessible via Dispatcher with correct cache headers
• [ ] JMX monitoring configured and showing metrics
• [ ] Log monitoring configured for replication errors
• [ ] Team runbook updated with agent details

Workflow 2: Production Incident Response

End-to-end workflow for diagnosing and resolving production replication issues.

Incident Triage

1. Gather Symptoms

**Questions to answer:**

• Is content replicating at all? (None vs. Some)
• Which agents are affected? (All vs. Specific)
• When did the issue start? (Timestamp)
• What changed recently? (Deployments, config, network)

**Data to collect:**

# Check agent status
curl -u $AEM_USER:$AEM_PASSWORD \
  http://localhost:4502/etc/replication/agents.author/<agent-name>.html

# Check queue depth
# Navigate to JMX Console: /system/console/jmx
# com.day.cq.replication:type=Agent,id=<agent-name>
# QueueNumEntries value

# Check recent errors
tail -n 100 <aem-install>/crx-quickstart/logs/error.log | grep -i replication

2. Diagnose Root Cause

**Delegate to:** [troubleshoot-replication](../troubleshoot-replication/SKILL.md)

**Follow diagnostic decision tree:** 1. Is queue blocked? → Network/connectivity issue 2. Are there 401/403 errors? → Authentication issue 3. Are there SSL errors? → Certificate issue 4. Is queue depth growing? → Target instance overloaded 5. Is content missing on Dispatcher? → Cache invalidation issue

**Common root causes:**

• Network partition between Author and Publish
• Service user credentials expired or revoked
• Target instance CPU/memory exhausted
• Dispatcher not accepting flush requests
• Firewall rule change blocking replication traffic

3. Implement Fix

**Based on diagnosis:**

**Network Issue:**

• Verify network connectivity: `ping publish-host`
• Check firewall rules
• Test replication port: `telnet publish-host 4503`

**Authentication Issue:**

• Verify service user exists and is active
• Check user permissions: `/useradmin`
• Regenerate credentials if expir