AI-supported server monitoring from gradwerk
Uptime, HealthCheck and WatchGuard analyses in one system
gradwerk develops AI-supported server monitoring software that not only measures, but also evaluates.
Traditional monitoring shows outages and curves, but often leaves teams alone with the categorisation. The gradwerk solution combines real-time checks with AI analyses that recognise patterns and summarise risks in an understandable way. This turns data into concrete information: What is conspicuous, why is it relevant and what to do next. The focus is on clear alerts, comprehensible reports and data protection in accordance with the GDPR.
Real-time monitoring that makes operations visible
The software monitors websites and servers at short, freely configurable intervals. A live dashboard shows status, response time, download speed and SSL status with expiry date. An availability history makes developments traceable over days, including incident tracking and automatic documentation. This allows teams to recognise not only "down" but also creeping deteriorations before users report them.
Alerts that inform instead of flooding
Alerts only work if they are reliable and do not end in alert fatigue. The alert system therefore works with threshold values, failed attempt counters and cool-downs to reduce false alarms and email spam. In addition to email, push notifications are available in the browser, supplemented by clear recovery messages as soon as a system is stable again. Acoustic notifications in the dashboard support a quick response during operation.
HealthCheck: AI classifies monitoring data and prioritises measures
The AI-based HealthCheck regularly analyses the monitoring data from the last few days and searches for anomalies and recurring patterns. Instead of raw data, the evaluation provides a comprehensible summary, a list of recognised anomalies and specific solution steps. A risk level assessment (e.g. Normal, Attention, Critical) prioritises the issues so that teams work first on the points that have a real impact on stability and performance. The system can optionally send the results as a report by e-mail.
WatchGuard and log analyses: keeping an eye on security signals
In addition to availability and performance, the software also focusses on security events. An AI-supported log analysis evaluates server logs and recognises typical attack patterns such as port scans, brute force attempts or unusual traffic peaks. A whitelist system reduces false positives by specifically excluding known patterns. The results appear as prioritised findings with instructions for action, so that "lots of log" quickly becomes a clear decision.
- Recognition of DDoS, brute force and scan patterns
- Risk level from normal to critical
- Whitelist rules against recurring false alarms
- Reports and direct links to analyses for quick coordination
Data protection and access: GDPR-first and clear roles
Monitoring processes technical data, so data protection is not an add-on. The software anonymises IP addresses during normal operation, filters sensitive parameters and uses staggered retention periods depending on the type of event. Roles and rights separate administration, analysis and pure view so that teams have targeted access.
What exactly does "AI-supported monitoring" mean in this software?
The software collects classic monitoring values such as accessibility, response time, download speed and SSL status. The AI also analyses this data at regular intervals and looks for deviations, trends and recurring patterns. This results in a comprehensible assessment with a risk level and prioritised recommendations. The goal is categorisation and the ability to act instead of pure data collection.
What is the difference between HealthCheck and WatchGuard analysis?
HealthCheck focuses on stability and performance via the monitoring history, for example on sudden slowdowns or recurring downtime patterns. The WatchGuard analysis focusses on log data and security signals, such as conspicuous IPs, brute force attempts or port scans. Both analyses use risk levels and specific action steps. The system thus covers operation and security from two perspectives.
How does the system prevent too many alarms and false alarms?
Notifications can be controlled with a down-count threshold so that an alarm is only triggered after several failed attempts. A configurable cooldown reduces repeated notifications in the event of persistent faults. In addition, the whitelist system supports log analyses to exclude known false positives. Recovery messages provide clarity as to when a problem has actually been resolved.
Which reports and summaries are available?
The software creates regular email reports on request for health check and log analyses as well as immediate notifications in the event of incidents. A daily briefing summarises the status of all monitored systems in a compact format. Report thresholds control whether only certain risk levels are reported. Analyses can also be shared via a link in order to quickly coordinate decisions within the team.
How does the solution support GDPR-compliant monitoring?
In normal operation, the system anonymises IP addresses and reduces user agent information to what is necessary. Sensitive query parameters are filtered and the storage duration is staggered depending on the log type. More complete data is only available where a security incident requires a more detailed check. The solution also supports organisational requirements such as documentation and transparent settings. For AI-based evaluations, AI systems from the EU and Germany are used.
Who is the software particularly suitable for?
Web agencies use it to monitor customer websites centrally and deliver comprehensible reports. IT departments receive a combined view of availability, performance and security events. Managed service providers benefit from scalable monitoring structures and automated analyses. In all cases, AI helps to set priorities and shorten response times.