Good news everyone
SMON is deprecated and will be removed soon.
We recommend moving your monitoring to RMON.
We recommend moving your monitoring to RMON.
v9.2.0 (planned: 15.11.2026)
- Redesigned the configuration viewer with collapsible sections, full-source view, search, match navigation, copying, and line wrapping.
- Added support for multiple HAProxy configuration files, including file creation, editing, saved versions, and Change Center integration.
- Added automatic detection of HAProxy configuration sources from systemd or Docker startup settings, with setup instructions when the service is not configured for multiple files.
- Added validation against the complete HAProxy configuration before applying file changes, with rollback if reload or restart fails.
- Unified service settings dialogs and configuration path handling across supported services, and fixed settings save failures caused by concurrent SQLite writes.
- Added a WAF Config Wizard for Enterprise and higher plans to create rules and exclusions, combine conditions for URI, IP, HTTP method, headers and request parameters, and preview the generated configuration.
- Added WAF rule synchronization from a Master to its directly linked Slaves in the same group, including rule files, catalog entries, Include state, edits, enable/disable actions and deletion. Conflicting replica rules and failed targets are reported; partial synchronization keeps changes pending.
- Made WAF rule creation and enable/disable actions safe to retry after interrupted requests, without duplicate files, Include directives, or database entries. Conflicting changes are reported instead of overwriting existing configuration.
- Made WAF mode changes atomic and confirmed the saved remote value for On, Off, and DetectionOnly modes. Changing the saved mode does not apply it to the running service.
- Added per-server and per-service WAF indicators for unknown, pending, and confirmed apply states. The Master's indicator includes confirmation from its linked Slaves. These track changes made through Roxy-WI; manual edits and service actions outside the WAF page are not tracked.
- Added Restart WAF and Reload NGINX actions to apply saved WAF configuration and confirm application of tracked changes. Failed or unconfirmed actions keep changes pending.
- Added full NGINX configuration validation when saving WAF rules, with restoration of the previous rule if validation or reload fails.
- Added HAProxy WAF agent readiness checks, including a SPOA protocol response, after restarting with an edited rule. Failed restarts trigger a rollback attempt, and recovery failures are reported.
- Fixed WAF editor read and save errors, service-specific runtime actions, and temporary-file cleanup. Lost remote responses no longer produce a successful-save indication.
- Unified the WAF overview with service cards, search, status filters and cluster information. Added WAF diagnostics with localized explanations and recovery guidance.
- Separated WAF event and audit logs in the shared log viewer, with complete audit transactions, search, Live mode and automatic log rotation.
- Fixed NGINX ModSecurity installation and configuration switching, preserving existing configuration and restoring it after failed validation or reload. HAProxy WAF installation now validates all active configuration files.
- Localized the revised WAF editor, rule creation, mode controls, and apply workflow in all six supported languages. Preserved compatibility with legacy WAF paths and Python 3.6 remote helpers.
- Added clickable worker statuses with heartbeat details, error explanations, troubleshooting guidance, and links to the affected server's settings.
- Grouped workers by instance and retained process run history, including previous starts and parallel processes, without counting old runs as additional active instances.
- Fixed Checker and Metrics controls to use active remote-worker availability instead of requiring their packages on the Web host.
- Fixed Overview chart refresh errors, oversized log tiles, clipped Change Center timeline icons, and navigation to server settings.
- Fixed mobile menu navigation, search, submenus and behavior after resizing. Simplified the recent-log panel on Overview with refresh and expandable content.
- Added localized HTTP error pages and replaced raw unexpected exception details with safe error messages. API and background requests retain JSON responses and HTTP status codes.
- Fixed HAProxy Statistics loading by using GET to display the page and POST only for submitted actions.
- Fixed database readiness monitoring at Web worker startup and preserved operator-supplied Gunicorn configuration.
- Added the Certificates catalog with separate disk and HAProxy runtime states, file upload, viewing and guarded deletion.
- Added queued HAProxy certificate replacement and upload into existing crt-list or certificate-directory binds without reload, with persistent files, full configuration validation and recovery across linked HA targets.
- Added per-server certificate History with pagination, target results and Resume recovery for incomplete rollback or finalization. NGINX and Apache retain file-management workflows.
- Updated Let's Encrypt deployment to use Runtime API for certificates already loaded in HAProxy, retaining the verified reload path for first deployment.
- Fixed migration of Let's Encrypt records whose original server has been deleted.
- Deprecated SMON; it will be removed soon. We recommend moving monitoring to RMON.
v9.1.0 (5.10.2026)
- Prepared Roxy-WI to run with the same application architecture from Linux packages, Docker Compose or Kubernetes.
- Added environment-variable overrides while preserving fallback to the existing
/etc/roxy-wi/roxy-wi.cfgfile and database settings. - Split the Roxy-WI runtime into independently managed Web, Scheduler, Service Events and Operations processes.
- Added a production Gunicorn launcher and separate commands for database migration, database readiness checks and every Roxy-WI process.
- Added liveness and readiness endpoints for container health checks and Kubernetes probes.
- Added a non-root production Docker image and Docker Compose configurations for SQLite and MariaDB installations.
- Published the Helm chart on GHCR with separate Deployments, a migration Job, Secrets, persistent storage, Ingress and configurable replica counts. See the OCI installation guide.
- Added safe defaults for read-only container filesystems and persistent Ansible data, configuration backups and runtime state under
/var/lib/roxy-wi. - Added support for reverse-proxy headers and a configurable public URL for Compose and Kubernetes installations.
- Moved service installation and other long-running Ansible work to durable database-backed operations delivered through RabbitMQ.
- Added operation leases, retry-safe execution, encrypted operation payloads, dead-letter handling and recovery of interrupted operations.
- Moved Ansible temporary files, SSH control sockets, inventories, downloaded Galaxy roles and collections to writable persistent paths.
- Improved installation failures so the actual Ansible error is stored and shown to the user instead of a generic web-server log message.
- Reworked Checker 5, Metrics 4, Port Scanner 2 and Socket 2 as independent long-running services connected through RabbitMQ.
- Added desired-state reconciliation, durable assignments, leases and worker heartbeats so distributed services recover their state after a restart.
- Added runtime version reporting by the new services without relying on the installed package version.
- Updated Overview and Admin Tools to show current process and worker status, replica counts, runtime versions, assignments and stale heartbeats without the old master-worker terminology.
- Kept systemd lifecycle controls for package-managed Roxy-WI processes and made container-managed processes explicitly controlled by Compose or Kubernetes.
- Added per-group deployment policies for each service: direct deployment, Change Center required, or both. Consolidated Admin area row actions into Actions menus.
- Added a local HTTPS quick start with generated persistent secrets, migration checks and an isolated Compose project.
- Moved filesystem and S3 backups from cron to Scheduler and Operations, with saved timezones, next-run status, bounded retries, SFTP atomic replacement and Boto3 transfers.
- Added migration commands for existing backup and certificate cron jobs, and automatic cleanup of completed operation history.
- Reworked Let's Encrypt around drafts, staging setup checks, reusable encrypted DNS profiles, automatic renewal, rate-limit handling and expiry alerts.
- Added atomic HAProxy certificate deployment, Runtime API verification and persistent rollback/recovery across HA targets.
- Moved Change Center actions to durable Operations tasks with visible task status, interrupted-worker recovery and a maintenance-window check when queued deployment starts.
- Added a structured log viewer and Live mode for internal and service logs, time ranges, literal search, expandable records, rotation handling and group access filtering.
- Added a shared container log journal and refreshed the Overview log panel; fresh HAProxy and NGINX installations now use JSON traffic logs while existing configurations are preserved.
- Added automated DEB and EL9/EL10 RPM publication and repository checks; improved database bootstrap, process health checks, notification delivery and SQLite write behavior.
- Hardened saved-configuration ownership, group-local map paths, certificate and map permissions, WAF temporary files, rendering of untrusted labels, and Fail2Ban authentication events.
v9.0.0 (05.09.2026)
- Added OpenID Connect authentication.
- Added a centralized workflow for managing HAProxy, NGINX, Apache and Keepalived configuration changes.
- Review syntax-highlighted configuration differences before deployment.
- Validate, approve, deploy, retry, cancel, and roll back changes from one place.
- Track every change with its author, target server, service, status, and timestamp.
- Added clear status indicators and improved deployment error messages.
- Added recovery controls for interrupted validation, deployment, and rollback operations.
- Canary deployment to selected nodes
- Configurable deployment batches and concurrency limits
- Manual promotion to the next node or batch
- Pause and resume controls
- Per-node retry and rollback actions
- Configurable health checks between deployment stages
- Ability to temporarily exclude unavailable nodes
- Scheduled deployments and maintenance windows
- Real-time deployment timeline
- Notifications through supported integrations
- Webhooks for external automation systems
- Continuous configuration drift detection
- Deployment duration and success-rate statistics
- Searchable audit history with detailed per-node results
- Exportable deployment reports
- Added confirmation dialogs for critical actions.
- Reworked local and remote command execution to prevent shell command injection.
- HAProxy Runtime API commands now use a direct TCP client instead of shell and netcat.
- Added input validation, timeouts and response-size limits for Runtime API operations.
- Hardened Nettools, Port Scanner, NGINX map generation and HAProxy configuration operations.
- Ansible inventory files now use unique names, restricted permissions and are always removed after deployment.
- Service deployment requests now verify access to every selected server.
- Credentials are no longer passed through shell commands when checking Apache status.
- Configuration comparison and HAProxy configuration editing no longer invoke shell utilities with user-controlled values.
- Port scans now run once with an explicit timeout and return a clear timeout error instead of causing a reverse-proxy 504.
- Services are marked as installed only after a successful Ansible deployment.
- Failed installation tasks are no longer overwritten with a completed status.
- Improved HAProxy Runtime API operations for servers, stick tables, ACLs and maxconn.
- Improved handling of JSON errors for AJAX requests.
- Authentication errors from background requests no longer redirect unexpectedly to the login page.
- Improved frontend error reporting when the backend response does not contain JSON.
- Fixed false 403 responses when a super administrator accesses servers from another group. Group administrators remain restricted to their active group.
- Fixed the HAProxy configuration versions page requiring an unnecessary server address.
- Fixed UDP listeners referencing a deleted HA cluster or server. The interface now shows a clear configuration warning instead of logging a traceback.
- Fixed layout and polling for UDP listeners with broken references.
- Fixed Apache status parsing and connection handling.
- Fixed nested task IDs returned after HA cluster deployment.
- Deleting users or groups now also removes related OIDC identities and mappings.
v8.3.0 (15.08.2026)
- A new map for HAProxy workflow
- Add a map for NGINX workflow
- A lot of security issues have been fixed
- A lot of bugs have been fixed
v8.2.6.5 (15.05.2026)
- A lot of security issues have been fixed
v8.2.6.3 (15.02.2026)
- A vulnerability in working with config comparing has been fixed
v8.2.6.2 (15.01.2026)
- A vulnerability in working with server syslogs has been fixed
- Improved error output when working with logs
v8.2.6 (15.12.2025)
- A new version of HAProxy has been added to the installation
- A new version of Node exporter has been added to the installation
- A new version of Apache exporter has been added to the installation
v8.2.5 (15.11.2025)
- A new version of NGINX exporter has been added to the installation
- The possibility to add NGINX cache
v8.2.4 (15.10.2025)
- The possibility to enable/disable websocket options for NGINX
v8.2.3 (15.09.2025)
- The possibility to hide/show NGINX backend server headers
- New UDP listeners are automatically added to the UDP listeners page
v8.2.2 (15.08.2025)
- The possibility to enable/disable HSTS for NGINX configs
- The possibility to hide/show server tokens for NGINX config
- The possibility to enable/disable security headers for NGINX config
- The possibility to add HTTP/2 in HAProxy bind option
v8.2.1 (15.07.2025)
- Asynchronous installation of services. Now it is not necessary to wait on the page for installation, a notification will come as soon as the service is installed
- New HA clusters are automatically added to the HA cluster page
- Showing the status of an HA cluster service has been added on the HA cluster page
- The possibility to add HTTP/2 in NGINX bind option
- The possibility to add name alias on the Add Nginx page
- Spanish language has been added
- AI assistance in editing configuration
v8.2.0 (15.06.2025)
- The possibility to create a server (Frontend) for NGINX vi the Add page
- Refactoring upstream creating for NGINX
- Roxy-WI refactored. Logs in JSON format now
- Database migrations have been added
- Error handling has been refactored
- Chinese language has been added
v8.1.8 (15.05.2025)
- HA Cluster status is updated automatically now
- Refactoring SSL page
- The possibility to download key and crt files
- Fixed and improve syntax highlights for WAF
v8.1.7 (15.04.2025)
- A new "log-forward" section and "mode log" for HAProxy have been added
- A new version of Node exporter has been added to installation
- Refactoring Roxy-WI log
v8.1.6 (15.03.2025)
- Checker services start to check the status of UDP listeners
- SSH fingerprint is deleted automatically now
v8.1.5 (15.02.2025)
- Roxy-WI shows backend server status for UDP listeners on the UDP balancer page
- Become master is updated automatically on the Keepalived overview page
- Status for HA cluster has been added
v8.1.4 (02.01.2024)
- Haproxy v3.1 has been added to installation on EL
- SSH keys are stored in the database
- A new version NGINX exporter has been added to installation
v8.1.3 (15.12.2024)
- Current connections are updated automatically on the NGINX overview page
- API endpoints and Terraform resource has been added for managing White and Black lists
- Dark them has been added
v8.1.2 (02.12.2024)
- The IP calculator has been added to the Nettools
- Requests per second are updated automatically on the Apache overview page
- Optimized working of Metrics and Checker tools
v8.1.1 (15.11.2024)
- Refactoring Let's Encrypt installation
- The possibility to use DNS API for issue Let's Encrypt certificates
- New way to manage Let's Encrypt certificates
v8.1.0.1 (21.10.2024)
- Bugs fix
v8.1 (16.10.2024)
- A new way to work with HAProxy config: now, when you add HAProxy section via the Add page, Roxy-WI keeps it. And you can edit your sections afterward via wizard.
v8.0.2 (02.10.2024)
- The possibility to mark SSH credentials as Shared to share them between groups
- Haproxy v3 has been added to installation on EL
v8.0.1 (15.09.2024)
- The possibility to enable/disable backend servers checks for UDP listeners
- The possibility to set parameters for backend server checks on UDP listeners
- New versions of NGINX and Node exporters have been added
v8 (02.09.2024)
- New API
- Transition to a new authorization system. Now JWT token is used instead of sessions
- Moving from Python 3.6 to Python 3.11
- Removing "/app" from URLs
- Information about UDP listeners automatically updated
- Servers on the Server page update their status and information automatically
- Action history for UDP listeners
- The possibility to set load balancing algorithm for UDP listeners
v7.3.2 (15.07.2024)
- Refactoring Filesystem backup installation
- Refactoring working with Git
- The possibility to view UDP balancer status
v7.3.1 (02.07.2024)
- The possibility to use VIP address as outgoing instead of server's IP
- Incoming and outgoing connections are updated automatically on the HAProxy overview page
- Refactoring S3 backup installation
v7.3.0 (15.06.2024)
- The possibility to create UDP load balancer
v7.2.6 (02.06.2024)
- Refactoring Geo installation
- Refactoring Grafana installation
- Refactoring Prometheus server installation
v7.2.5 (15.05.2024)
- The Settings page is AJAX page now
- Metrics began to be updated in real time
- The Server page has gone, only the Admin area left
v7.2.4 (02.05.2024)
- The possibility to reconfigure SMON Agent
- The possibility to send alerts via MatterMost
- The port is taken directly from settings for SMON service now. Without editing service config
v7.2.3 (15.04.2024)
- The Whois tool has been added to the Nettools
- The possibility to clone settings for S3 backup
- Refactoring WAF installation
v7.2.2 (02.04.2024)
- The Backup page is AJAX page now
- Network status has been added to the System Info
- The channels for alerting have separated page
- A new version of HAProxy has been added for installation
v7.2.1 (15.03.2024)
- Roxy-WI checks if actions (start, stop, reload, restart) with services was done properly
- Roxy-WI has started to return error after service actions (start, stop, reload, restart)
- The possibility to check config for Keepalived service
- Actions history for HA clusters has been added
v7.2.0 (02.03.2024)
- SMON now has agents. You can create multiple agents in different DCs, regions, etc. and control your endpoints from there.
- SMON Admin page has been deleted. Now all actions with checks on SMON Dashboard page
- The possibility to set check interval for every check
- Full SMON refactoring: it works faster and consumes less resources
v7.1.2 (15.02.2024)
- Refactoring HAProxy, NGINX, Apache, Keepalived and Node exporters installation
- The possibility to add SSH keys with passphrase
- SSH passwords and SSH passphrases are crypted now. Do not forget change a new option "secret_phrase" in the config
v7.1.1 (02.02.2024)
- The possibility to install Apache via HA cluster
- New way to show Master-Slave relationship and HA cluster on the Overview pages
- New versions of Prometheus exporters have been added
v7.1.0 (15.01.2024)
- A new entity like HA cluster has been added
- A new way to work with Keepalived service - HA cluster
- Refactoring HAProxy, NGINX, Apache and Keepalived installation
- The bugs with Port scanner and SMON in Docker has been fixed