Features
Everything Replication Manager does for MySQL, MariaDB and PostgreSQL in production, and what Cloud18 adds around it.
Replication Manager
Everything it does, by area
Monitoring, high availability, schema, security, logs, alerting, the technical base, maintenance, the orchestrator, the configurator and the workload.
Simplicity
- A single binary: the server, the web interface, the API and the command line
- Embedded metrics: its own Carbon/Graphite store, nothing else to install
- Scripts and templates embedded in the binary
- REST API documented with Swagger
- MCP server for your AI assistant
- Scripting hooks on every event: before and after a failover, backups, schema and variable changes, alerts
- API tokens, each with its own grants
- Signed plugins that extend monitoring, security and archiving
- Built-in regression tests (regtest): failover, switchover, rejoin and backup scenarios, run against your own cluster
Monitoring
- Topology, replication lag and replica state
- Slow queries, query digest and Performance Schema: mutexes, latches, memory
- Long and sleeping transactions that hold locks
- Graphs: workload, replication, resources, InnoDB, memory
- The internal network of every database, proxy and app
- Resource usage per service: CPU, memory, IO, disk
- Live progress of every running backup
High availability
- Automatic failover and controlled switchover
- A switchover waits for long write transactions, it never kills them
- Arbitration across data centers, by majority
- Rejoin of the former primary: flashback, pg_rewind or reseed from a backup
- GTID and semi-synchronous replication against data loss
- Several replication-managers on one cluster: one active, the others standby
Staging, child clusters and CDC
- Staging clusters: your developers test on production data, seeded from the parent and refreshed on demand
- Child clusters: test a migration, another release or another storage engine, under the parent’s real writes
- Created in one call from the GUI, the API or the MCP server, the plan checked first
- Seeded and reseeded from the parent’s backup, then replicating from it
- CDC replication plugins: supervised processes that carry a source’s changes to another engine, MariaDB ColumnStore first
Schema
- Tables, sizes and indexes of every database, scanned daily
- Schema events: every table created, altered or dropped, with its column diff
- Advisor plugins: duplicate indexes, auto-increment exhaustion, row size, LOB compression
- Table checksums between the primary and its replicas
- A hook on schema changes for your own scripts
Security
- Security score from 0 to 100, grade A to F, 19 checks on every server
- Detection plugins: privilege escalation, off-hours access, clear-text passwords and card numbers in binlogs
- Users and grants audit, password rotation
- Secrets kept in Vault
- TLS and per-user API grants
- Single sign-on with the Cloud18 identity
- Signed plugins and a software bill of materials
Logs
- Error log, slow query log and audit log of every server, collected and shown in the GUI
- SQL errors tailed as they happen
- A log level per module: backup, proxy, heartbeat, arbitration, scheduler and more
- The cluster state log: every decision of replication-manager, explained
- Logs read from the GUI, the API and the MCP server
Alerting
- Health alerts on every database and proxy
- Email, Slack, Microsoft Teams and Pushover
- Your own alert script on every event
- Warnings that name the cause, for example the server-id colliding on a binlog stream
- Detection plugins raise their own alerts
Technical
- Topologies: master and replicas, multi-tier, multi-source, delayed, MariaDB Galera, MySQL group replication, active/passive, PostgreSQL WAL streaming and logical
- Proxies: HAProxy with its runtime API, ProxySQL, MaxScale
- REST API, command line and MCP server
- MariaDB up to 12, MySQL and Percona Server 8.4, PostgreSQL 17
Maintenance
- Logical and physical backups: mysqldump, mydumper, dumpling, splitdump, mariadb-backup, xtrabackup, pg_dumpall, pg_basebackup
- Encryption at rest, archives on S3, Proxmox Backup or Plakar
- Point-in-time recovery from binlog and WAL copies; partial restore of one table or one schema
- Scheduler: backups, optimize and analyze, binlog purge
- Password rotation
- Configuration kept in git
Orchestrator
- OpenSVC, Kubernetes or on premise
- Databases, proxies and applications provisioned and restarted by replication-manager
- Failover volumes for active/passive services
- Rolling restart, rolling upgrade and rolling reprovision
- Dynamic resize of CPU, memory, IO and disk, inside the plan and up to your limit
Configurator
- Every server sized from its plan: memory, cores, IO
- Rulesets that tune MariaDB, MySQL and PostgreSQL to that size
- Tags for the workload and the features you want
- Configuration changes tracked, with the delta of every file
- The database images of your own list, for every upgrade
Workload
- Top queries and the query digest, per server
- A benchmark in one click: sysbench through the proxy, from replication-manager
- Traffic simulation through the proxies, to prove the write path
- Workload tags that adapt the configurator to your use
- Detection plugins: slow query regressions, full table scan spikes, connection, error and temporary table storms, metadata lock contention
- A replication lag predictor
Cloud18
What Cloud18 adds
Applications next to the data, and a marketplace of partners.
Applications next to the data
- A template library: ERPNext, Dolibarr, Forgejo, phpMyAdmin, Adminer, Valkey, Memcached, RabbitMQ, MinIO, RustFS
- Your own code in PHP, Java, Next.js, Python, Go or .NET, cloned from git
- A schema and a user created for each app
- Start, stop, restart and resize from the GUI, the API or the MCP server
The marketplace
- A welcome credit of 1000 € on the infrastructure for every new account
- Units: DBU, APU, BKU, BAU and GWU
- The same units at every partner, each at its own price
- Guaranteed resources; beyond the plan, up to your limit, at the over-commit price set by each partner
- A monthly unit statement per cluster
- Self-service clusters
- Support, identity, external backups and knowledge base
- Support by chat, right in the application, and our experts’ services
- A cloud storage zone for your backups and archives
- Your configurations backed up in the cloud
- A knowledge base on configurations, security and bugs, and the documentation
Next steps
Install Replication Manager on premise, or start on Cloud18.