Overview

5.1.1 Monitoring Overview

replication-manager runs a continuous monitoring loop that tracks every server in every managed cluster. On each tick (default every 2 seconds, controlled by monitoring-ticker) it connects to each database and proxy server, reads state, and updates an in-memory topology model that drives the GUI, API, alerting, and failover decisions.

5.1.2 What the Monitor Tracks

5.1.2.1 Replication topology

• Server role (primary / replica / candidate / down)
• GTID positions and replication lag
• Semi-sync acknowledgement state
• Slave SQL/IO thread status
• Relay log position and relay log size
• Replication domain IDs (multi-source)

5.1.2.2 Server health

• Global status variables (connections, InnoDB metrics, query cache, etc.)
• Global variables (for config drift detection)
• InnoDB engine status
• Error log events
• Schema changes (DDL detection — requires monitoring-schema-change)

5.1.2.3 Workload

• Active processlist (up to 50 longest-running queries — monitoring-processlist)
• Performance Schema digest statistics (monitoring-performance-schema)
• Query routing rules consolidated from all proxies (monitoring-query-rules)

5.1.2.4 Proxies

• ProxySQL, HAProxy, and MaxScale connection counts and routing state
• Backend server weights and online/offline transitions

5.1.2.5 Metrics

• All collected counters are pushed to the embedded Graphite database and exposed as Prometheus metrics
• See Metrics for Graphite and Prometheus configuration

5.1.3 Monitoring Loop

Every monitoring-ticker seconds:
  for each cluster:
    for each server (parallel goroutines):
      connect → read status → update topology model
    evaluate topology → compute failover eligibility
    fire alerts if state changed (see Alerting chapter)
    push metrics to Graphite
    write API state

When monitoring-capture is enabled, replication-manager automatically saves a diagnostic snapshot (SHOW SLAVE STATUS, SHOW PROCESSLIST, SHOW GLOBAL STATUS, SHOW ENGINE INNODB STATUS) for a configurable number of ticks whenever a monitored error code appears. See Logs and Troubleshooting for details.

5.1.4 Accessing Monitoring Data

5.1.4.1 GUI

The web dashboard shows the live topology, server state, processlist, and alerts. The React dashboard is served at the replication-manager API address (default port 10001).

5.1.4.2 REST API

GET /api/clusters/{clusterName}/topology/servers
GET /api/clusters/{clusterName}/topology/master
GET /api/clusters/{clusterName}/topology/slaves
GET /api/clusters/{clusterName}/topology/alerts
GET /api/clusters/{clusterName}/topology/logs

5.1.4.3 Prometheus

GET /api/prometheus

Returns all collected metrics in Prometheus text exposition format. Point your Prometheus scraper at this endpoint.

5.1.4.4 CLI

replication-manager-cli status --cluster=my-cluster
replication-manager-cli status --cluster=my-cluster --with-errors

5.1.5 Configuration