Changelog

Absolute fault tolerance: We have completely redesigned the architecture of the monitoring core, separating the processes of storing analytics and performing pings. Now all subsystems work completely autonomously. Even during peak loads on servers, scheduled database maintenance, or network anomalies, your site checks continue without the slightest delay.

Instant start and smart distribution: A new algorithm for "lazy" loading of checks (Lazy Loading) has been introduced. The service is capable of scaling in fractions of a second, smoothly distributing the load and eliminating the creation of "bottlenecks" even while simultaneously serving hundreds of thousands of resources.

Guarantee of metric preservation: The internal log buffering mechanism has been rewritten. A new hardware-independent queue guarantees 100% preservation of the history of checks and incidents.

Modular backend architecture: We conducted a deep refactoring of the codebase. We divided large monolithic processes into independent micro-components and completely isolated the database query building layer. Visually, the interface remains the same, but this solid foundation multiplies overall system performance and allows us to roll out new features significantly faster, eliminating the risk of disrupting current monitoring stability.

HTTP Integration (Webhooks): Now you can route incident notifications directly to your servers. We send a clean and structured JSON payload upon every status change. This opens up endless possibilities for integration with your internal dashboards, or custom auto-restart scripts.

Security (HMAC-SHA256): We took your data protection seriously. Every outgoing webhook is automatically signed with a unique 256-bit cryptographic key. You can always be 100% sure that the incoming request came exactly from our service, and not from malicious actors.

We have completed a massive localization update! Now, a fully-fledged multi-language system is available not only on public pages but also inside the control panel (on the .app subdomain).

The entire service interface is completely translated and adapts to the selected language, making monitoring setup and reading analytics maximally comfortable for users from all over the world.

We have added a fully-fledged localization system! Language switching is now available on all public pages of the site, making it even more convenient for our global audience to interact with our service.

Note: for internal control panel pages (on the .app subdomain), multi-language support is in active development and will be added in upcoming updates.

A new "Incidents" tab in the main menu! You no longer need to go into the settings of each individual monitor to check its downtime history. We have combined all failures across all your resources into a single global dashboard.

All failure history is available here with convenient filtering, sorting, and quick navigation to the details of a specific problematic service.

Calculation core: We have completely redesigned the time calculation architecture in the "Uptime" block. The uptime metric now works with 100% accuracy and has become completely independent of the incident history. Even with the automatic clearing of old failures or log archiving, the counter will remain crystal clear and accurate.

Extreme performance: The delivery of system events to the browser has been significantly optimized. Thanks to the transition to a single data channel for the entire account, the interface consumes significantly less memory and works lightning fast.

Smart history interface: The list update logic has been refined. If you are examining old records on distant archive pages, a sudden new failure will not "shift" your data — the system will gently notify you of the new event while preserving your current position on the page.

We have improved the deep analytics interface to make reading statistics even faster and more comfortable:

• Charts now adapt better to high latency spikes while maintaining the clarity of everyday response metrics.
• Tooltips have become neater and are now automatically highlighted in color during downtimes.

Strict control of timeouts and dead DNS: A multi-level protection mechanism against "hung" servers has been implemented. If the checked site intentionally or due to a failure holds the connection endlessly, the scanner will safely and guaranteed terminate the session. You will see a fair timeout failure status, and our statistics will remain crystal accurate.

Strict limit on parallel connections: We have trained the internal balancer to control the number of simultaneous live requests to a single site. This 100% eliminates the risk of accidentally creating a DDoS effect on your servers from our side, even if they are operating unstably and take a long time to respond.

We have completely rewritten the internal monitoring network core to work under extreme loads. The system now operates completely asynchronously, instantly distributing thousands of checks without the slightest delay in queues.

Work with connection pools and RAM has been radically optimized — the service has become even faster and more stable at digesting the check of a large number of domains without performance loss.

We have completely reworked deep analytics: made metrics more accurate and charts more transparent. Instead of smoothing, a stepped display is used, reflecting the real dynamics of the response. We updated the tooltips with details on checks, errors, and delays, and added statistics below the chart: uptime percentage, average response, median, maximum, and P95.

We also optimized loading: the dashboard opens instantly, and heavy analytical calculations are performed asynchronously in the background.

Instead of obscure system codes (like econnrefused or nxdomain), the interface will now tell you exactly what the problem is (e.g., Nginx/Apache went down, invalid DNS settings, or the server forcibly reset the connection).

Navigation menus (both on the main site and in the control panel) now open fullscreen, stylishly block background scrolling, and feel like a native application.

Solved a complex layer overlapping issue in the monitor list. Dropdown option menus now consistently overlay the floating monitor creation button and adjacent cards.

Fixed a bug where opening a monitor menu on a smartphone didn't close old cards, and the transparent background let through accidental clicks on other interface elements.

Launched a full-fledged system for managing your alerts. Now you can set up integrations, link multiple chats, and temporarily pause any notifications in one click.

Telegram messages about resource recovery (UP) now display the real server response time (ping) in milliseconds.

The internal logic for sending notifications has been completely redesigned. The new queue processing architecture has made the system more reliable, laying a solid foundation for the quick implementation of new channel types (Webhooks, Email, Slack, Discord).

We redesigned the interface for creating and editing monitors. We dropped hidden tabs in favor of a single page where all parameters are grouped into convenient logical blocks.

Significantly increased the fault tolerance of background queue workers. Added elegant handling of business exceptions, which prevents task looping, saves database resources, and keeps system logs clean.

Implemented a new two-level security contour for recording analytics. Fixed a rare situation where deleting one monitor could briefly block the saving of metrics for other resources. Now your uptime charts and statistics are fully protected against data loss.

We completely separated the public site and the control panel. Now the entire application works in an isolated and secure environment on the app. subdomain. This increases stability, speeds up the interface, and makes the architecture truly scalable.

We also prepared our public storefront for SEO indexing while securely closing private dashboards from being scanned by search robots.

Fixed mobile navigation flaws. When clicking on links, the mobile menu now smoothly closes, instantly sending you to the required section of the landing page without overlapping the content.

Significantly increased the fault tolerance of the monitoring core during peak network loads. Delays and timeouts during mass notification mailings have been eliminated, and the handling of non-standard responses and network anomalies of monitored sites has been improved.

We added basic Telegram integration! Now you can receive notifications about service downtimes right in the messenger.

To avoid disturbing you with false positives due to momentary network blinks, we made alerts "smart". At the first error, the system does not sound the alarm immediately but does a series of instant rechecks. A notification is sent only if the server is truly unavailable.

As a bonus, we added context: the history of the last 10 checks before the incident is also attached to the error message, so you can immediately see if the response time was growing before the failure.