Sleeping Transactions Monitoring: Monitor long-running transactions including sleeping transactions that hold InnoDB locks. Tracks transaction duration, isolation level, tables locked, rows modified, and memory usage. See documentation
Performance Schema Mutex Monitoring: Collect InnoDB mutex wait events from Performance Schema. Identifies mutex contention and performance bottlenecks. See documentation
Performance Schema Latch Monitoring: Monitor InnoDB read-write lock (rwlock) wait events. Diagnoses concurrency issues with read/write operations. See documentation
Performance Schema Memory Monitoring: Track memory consumption by InnoDB subsystems via Performance Schema memory instrumentation. See documentation
Per-Module Log Levels: Configure log verbosity independently for each subsystem (backup, proxy, heartbeat, scheduler, etc.). 20+ new log-level-* parameters for fine-grained logging control
Schema Change Script Hook: New monitoring-schema-change-script called when a table is created, altered, or dropped. Receives a column diff on stdin showing exactly what changed. See documentation
Variable Change Script Hook: New monitoring-variable-change + monitoring-variable-change-script detects when a server variable changes over time (e.g. SET GLOBAL). Pipes a before/after diff to stdin. Distinct from monitoring-variable-diff which compares master vs slaves. See documentation
Config Tracking Fixes: Prevent empty dummy.cnf overwrite on repman shutdown. Delta computation skipped when server is down. New GUI Config Files panel with Clear Delta button.
On-Premise Rolling Upgrade: Upgrade scripts for Debian and RHEL with version pinning. Two-axis db_distributions.json for OS family × deploy method. See documentation
OpenSVC Instance-Level Start/Stop: Switch to per-node instance API to avoid 409 "warn state" errors. New opensvc-use-orchestrated-start flag for HA-safe abort+restart. See documentation
Server Menu: Restart and Upgrade: New "Restart Database" and "Upgrade Database" entries in the server context menu.
Docker Image Version Sort: Version dropdown sorts by semantic version descending (newest first).
Schema Change Script Hook: The monitoring-schema-change-script config key is now wired into the monitoring loop. Called when a table is created, altered, or dropped with a column diff piped to stdin showing exactly what changed. See documentation
Config Tracking Fixes: Prevent empty dummy.cnf overwrite on repman shutdown that caused all variables to show as "Deprecated (loose)". Delta computation skipped when server is down. New GUI Config Files panel with Clear Delta button. See documentation
On-Premise Rolling Upgrade: New upgrade scripts for Debian and RHEL with version pinning from prov-db-docker-img. Two-axis db_distributions.json for OS family × deploy method, pushed by back office. StopDatabaseServiceClean with innodb_fast_shutdown=0 for safe major version upgrades. See documentation
OpenSVC Instance-Level Start/Stop: Switch from orchestrated to per-node instance API to avoid 409 "warn state" errors. Optimistic clear before start. New opensvc-use-orchestrated-start flag for HA-safe abort+restart recovery. See documentation
Server Menu: Restart and Upgrade: New "Restart Database" (generic stop+start for all orchestrators) and "Upgrade Database" (clean stop + upgrade script) entries in the server context menu.
Docker Image Version Sort: Version dropdown in the configurator now sorts by semantic version descending (newest first) instead of alphabetically.
Settings Shortcut Buttons: Gear icon buttons on dashboard, maintenance, schema, tops, and graphs panels link directly to the relevant settings section. Clicking navigates to the Settings tab with only the related accordion visible; a "Show all settings" button restores the full view.
Schema Tab Rename: The "Shards" tab has been renamed to "Schema" to better reflect its purpose (table list, checksum, schema graph).
DictTables Persistence Separation: Table metadata (DictTables) is now saved to a separate dicttables.json file instead of serverstate.json. This prevents table size changes from generating git diffs on every monitoring tick. The file is automatically added to .gitignore. Cluster workload totals (DBTableSize/DBIndexSize) are recomputed immediately on restart from the cached metadata so the dashboard no longer shows "cluster totals missing" after a fresh restart.
Signup Email Verification Polling: After signup, the frontend polls a lightweight CRM endpoint (GET /api/signup/status?email=xxx) every 5 seconds to check GitLab email confirmation status. Once confirmed, the user is automatically logged in. This avoids hitting the login rate limiter (3 attempts / 3 minutes) that would block a polling-based login approach.
Restic Task Management: Background task queue for backup operations with pause/resume/cancel support via API
Automatic Disk Space Management: Restic automatically purges oldest backups when disk space is critically low (backup-restic-purge-oldest-on-disk-space)
Disk Threshold Configuration: Separate thresholds for warning and critical disk usage (backup-disk-threshold-warn, backup-disk-threshold-crit)
Restic Timeout Control: Configure operation timeouts for long-running backup/restore operations (backup-restic-timeout)
Local Repository Support: Use local filesystem repositories with per-cluster isolation (backup-restic-local-repository)
MariaDB Backup Tools: Native support for mariadb-backup and mariadb-import (formerly maria-backup)
Configurator Diff Tracking: Track differences between proposed and current configuration
Preserve Remote Configuration: Configurator maintains remote settings during updates
Deprecated Configuration Detection: Automatically identify outdated configuration parameters
OpenSVC Go Agent: Migration to new OpenSVC go-based agent
OpenSVC Service Logging: Enhanced logging for OpenSVC agent operations
OpenSVC Failover Monitoring: Split-brain detection and monitoring for OpenSVC clusters
Application Deployments: Support for managing application deployments via peering
Compact JSON Responses: Reduced API response sizes for improved performance
OAuth Authentication: OAuth2 provider integration for API authentication (parameters: api-oauth-client-id, api-oauth-client-secret, api-oauth-provider-url, api-oauth-redirect-url)
Enhanced Scheduler System: Comprehensive task scheduling with cron support for database maintenance, backups, and administrative operations
Dual-Mode Job Dispatch (sql/api): New scheduler-jobs-mode configuration. In api mode, tasks are dispatched via REST API cookies instead of polling the SQL jobs table, removing the need for the replication_manager_schema.jobs table on database hosts. The dbjobs script authenticates via JWT and checks task availability through the replication-manager API. Default mode remains sql for full backward compatibility. See documentation
Configurable Task Execution Mode: New scheduler-jobs-exec-remote parameter allows forcing specific tasks to run remotely via the dbjobs script instead of locally in replication-manager. Useful for tasks like mysqldump where the client binary version must match the database server version. See documentation
Embedded dbjobs Script: The dbjobs maintenance script is now embedded directly in the replication-manager binary via go:embed, bypassing the 65535-byte truncation limit of the OpenSVC compliance moduleset database column. Existing deployments receive the full script automatically via the built-in script auto-upgrade mechanism.
Job API Security: Job dispatch API endpoints now require the db-jobs ACL grant and JWT authentication. The system service account created by secret-login receives this grant automatically as part of the db grant group.
Schema Monitoring Default: Default cron for schema monitoring changed from once-a-year (0 0 1 1 * *) to daily at 2 AM (0 0 2 * * *). On startup, if no cached table metadata exists, a one-time schema scan is triggered automatically.
Optimize Cron Fix: Default optimize cron fixed from 0 0 3 1 * 5 (1st of month AND Friday — rare) to 0 0 3 1 * * (1st of every month at 3 AM).
Metadata Lock Avoidance: The jobs table creation now checks information_schema before issuing DDL, avoiding unnecessary CREATE DATABASE IF NOT EXISTS and CREATE TABLE IF NOT EXISTS statements that could hold metadata locks during backups.
Rolling Restart/Reprovision: Schedule automatic rolling restarts and reprovisioning operations
SLA Rotation: Automated SLA metric rotation via scheduler
MySQL 8.4 Support: Version parsing and SSL mode compatibility fixes for MySQL 8.4
PostgreSQL Enhancements: Improved PostgreSQL logical and streaming replication support
Job Script Versioning: Automatic upgrades and version tracking for database job scripts
Database Jobs Schema Upgrades: API endpoint for upgrading jobs schema
SQL Error Tailers: Monitor and tail SQL error logs for analysis
Audit Log Tailers: Capture and monitor audit log entries