Every minute an online shop is unreachable costs money and trust. One hour of downtime causes average revenue losses of 5,600 euros for mid-market e-commerce companies (Gartner, 2025) -- multiples of that for larger shops. Yet many operators learn about outages through customer complaints or declining revenue figures -- often hours after the shop has gone offline. Professional uptime monitoring detects outages within 30 seconds and automatically initiates countermeasures before the first customer notices an error.
Synthetic Monitoring: Proactive Availability Checks
Synthetic monitoring simulates real visitor behavior through automated HTTP requests to the shop. At regular intervals -- typically every 30 to 60 seconds -- the monitoring system sends requests to defined endpoints and checks status code, response time and page content. Unlike passive monitoring, synthetic monitoring does not wait for real visitors but tests proactively -- including nights and weekends when no internal team is watching the shop.
The key checkpoints for an online shop are: homepage (HTTP 200, load time under 2 seconds), category pages (product lists load correctly), product detail pages (prices and availability displayed), shopping cart (products can be added) and checkout entry (login or guest order page reachable). A shop that is reachable on the homepage but throws an error in checkout loses revenue -- without a simple HTTP check detecting it.
Professional synthetic monitoring therefore checks not only HTTP status but also page content: does the response contain expected keywords? Are prices displayed correctly? Is the contact form present? This content validation also detects partial outages -- for instance when the database is unreachable but the web server delivers an error page with HTTP 200. The monitoring setup defines individual check criteria for each endpoint.
HTTP Monitoring
Status code, response time and TLS certificate checks every 30-60 seconds from multiple locations.
Real-User Monitoring
Capture real user experiences: Core Web Vitals, page load time and errors in actual browsers.
Alerting Chains
Escalation levels: email instantly, SMS after 5 minutes, phone call after 15 minutes -- until someone responds.
Status Page
Public status page shows current availability of all systems and provides transparent updates during incidents.
SLA Tracking
Automatic calculation and reporting of availability against defined SLA targets (99.9%, 99.95%).
Multi-Location Checks
Checks from 3+ geographic locations prevent false positives from local network issues.
Real-User Monitoring: Measuring Actual Experience
Synthetic monitoring checks availability from a standardized test client's perspective. Real-User Monitoring (RUM) captures the actual experience of real visitors: How fast does the page load on a mobile device over 4G? What is the error rate for visitors with older browsers? Which pages cause the most JavaScript errors? RUM provides data no synthetic test can replicate.
RUM's core metrics are the Core Web Vitals: Largest Contentful Paint (LCP, loading speed), First Input Delay (FID, interactivity) and Cumulative Layout Shift (CLS, visual stability). Google uses these metrics as ranking factors -- a shop with poor Core Web Vitals loses customers through slow load times and visibility in search results. According to Google (2025), pages with good Core Web Vitals have a 24 percent (project experience) lower bounce rate than pages with poor scores.
The combination of synthetic monitoring and RUM provides a complete picture: synthetic monitoring detects outages in seconds, RUM shows actual user experience across all devices, browsers and connection speeds. Together they enable not only reacting to outages but also detecting performance degradations before they impact conversion. The professional monitoring setup integrates both approaches into a unified dashboard.
Alerting Chains: From Alarm to Action
Monitoring without alerting is a dashboard nobody watches. The value of monitoring is created through the alerting chain: automated notification of the right people through the right channel at the right time. A professional alerting chain is built with escalation -- if the first notification is not acknowledged, the next escalation level follows.
A proven escalation model: Level 1 -- email and push notification immediately upon outage detection. Level 2 -- SMS to the on-call technician after 5 minutes without acknowledgment. Level 3 -- phone call to the on-call service after 15 minutes without acknowledgment. Level 4 -- phone call to the responsible project manager after 30 minutes. This chain ensures no outage goes unnoticed -- not even at 3 AM or on weekends.
Equally important as escalation is preventing alert fatigue: too many notifications lead to alarms being ignored. Multi-location checks prevent false positives from local network issues -- an alarm is only triggered when at least 2 of 3 monitoring locations confirm the outage. Threshold-based alerts distinguish between warning (response time over 2 seconds) and critical alarm (HTTP 5xx or timeout), keeping the team focused on truly urgent problems.
The escalation logic is the most critical part of the alerting chain: if an alarm is not acknowledged within 15 minutes, it automatically escalates to the next level. After 30 minutes without response, the team lead is notified. After 60 minutes, the technical director. This automatic escalation ensures no alarm goes unanswered -- even at night, on weekends or during vacation periods. Escalation levels and timeout values must be regularly tested to ensure the monitoring infrastructure functions as planned in an emergency.
Status Pages: Transparency During Incidents
A public status page shows current availability of all systems and proactively informs during incidents. Instead of customers calling support to ask if the shop is down, they see the current status on the status page, estimated resolution time and progress updates. This transparency reduces support volume during incidents by up to 60 percent (project experience).
The status page typically shows individual system status (website, checkout, payment processing, API), uptime over the last 90 days as a timeline, planned maintenance windows and current plus past incidents with activity logs. It is automatically populated from the monitoring system and manually supplemented with comments. For professional shop maintenance, a status page is an indispensable communication tool.
Implementing a status page requires a clear communication strategy: what information is published when? Too much transparency can unsettle customers, too little transparency undermines trust. Best practice is a three-stage model: Detection ("We are investigating an issue"), Working ("The problem has been identified, we are working on a solution") and Resolution ("The problem has been resolved, all systems operating normally"). Each stage is documented with a timestamp so customers can track progress.
For online shops with Service Level Agreements, the status page additionally serves as a compliance instrument: it documents availability outages transparently and traceably. At month end, actual availability can be checked against the SLA target -- including the duration of each individual incident. This documentation is also valuable internally: it reveals patterns (frequent outages at specific times), identifies weak points and provides the data basis for infrastructure investments.
SLA Tracking and Availability Reports
Service Level Agreements (SLAs) define the guaranteed availability of an online shop. The most common SLA tiers are 99.9 percent (maximum 43.8 minutes downtime per month), 99.95 percent (maximum 21.9 minutes) and 99.99 percent (maximum 4.3 minutes). SLA tracking automatically calculates actual availability from monitoring data and compares it against the target.
| SLA Tier | Maximum Downtime per Month | Maximum Downtime per Year |
|---|---|---|
| 99.5% | 3.6 hours | 43.8 hours |
| 99.9% | 43.8 minutes | 8.76 hours |
| 99.95% | 21.9 minutes | 4.38 hours |
| 99.99% | 4.3 minutes | 52.6 minutes |
Monthly availability reports document uptime, number and duration of incidents, average response time and SLA compliance. These reports are valuable not only for internal reporting but also for communication with business partners and marketplace integrations that frequently require SLA documentation.
An often underestimated aspect is distinguishing between planned and unplanned downtime. Maintenance windows for updates are planned downtime and are reported separately in SLA tracking. Only unplanned outages count against the SLA target. This differentiation is important for fair availability assessment and planning maintenance windows.
DNS Monitoring and Performance Tracking
DNS issues are an often overlooked cause of outages. When the DNS server is unreachable or delivers incorrect entries, the shop is inaccessible to all visitors -- even though the web server itself is functioning perfectly. Professional DNS monitoring checks the resolution of all relevant DNS records (A, AAAA, CNAME, MX) at regular intervals and detects manipulations or outages within seconds.
Performance tracking adds a quantitative dimension to availability monitoring: how does the shop's response time develop over weeks and months? Gradual performance degradations -- from growing databases, increasing plugin load or inefficient caching configurations -- become visible before they noticeably impact user experience. According to an Akamai study (2025), every 100 milliseconds of additional load time leads to a conversion decline of 1 percent in e-commerce. Proactive performance tracking is therefore not just a technical measure but a direct lever for business success.
Multi-Location Monitoring: Checking Location-Dependent Availability
A single monitoring location only detects outages affecting that location. A CDN error, routing issue or regional DNS outage goes undetected if the monitoring server sits in the same data center as the shop. Multi-location monitoring checks reachability from multiple geographic locations -- ideally from at least three different regions (project experience). Only when multiple locations report an outage is an alert triggered. This significantly reduces false alarms from temporary network fluctuations.
For international online shops, multi-location monitoring is particularly relevant: customers in different countries reach the shop via different network paths, CDN endpoints and DNS resolvers. A shop may be reachable from Germany while customers in Austria or Switzerland see an error message. Professional monitoring setups check availability from all relevant markets and measure latency per location -- detecting not just total outages but also performance degradations in specific regions.
Evaluating Monitoring Data: Recognizing Trends and Patterns
Raw monitoring data only becomes actionable information through systematic evaluation. Trend analysis is the most important tool: if average response time slowly increases over weeks, this indicates growing database problems, increasing traffic without corresponding scaling or accumulating technical debt. These gradual deteriorations go unnoticed in daily operations but are clearly visible in monitoring trends.
Pattern analysis identifies recurring problems: if elevated response time occurs every Tuesday between 2pm and 4pm, this may correlate with a cron job, a backup run or a traffic peak from newsletter sending. These correlations are nearly impossible to recognize without monitoring data. Professional evaluation transforms monitoring data into concrete optimization measures and prevents small performance losses from becoming larger problems.
SSL and Certificate Monitoring
An expired SSL certificate makes an online shop inaccessible to customers -- modern browsers display a warning page that virtually no visitor bypasses. Yet expired certificates are among the most common preventable causes of outages. Certificate monitoring checks the validity of all SSL certificates and warns 30 days before expiration -- enough time to organize renewal, even when automatic renewal processes fail.
Beyond simple validity, professional SSL monitoring also checks the certificate chain (chain problems cause errors on certain devices), TLS version (TLS 1.0 and 1.1 are considered insecure and rejected by browsers), cipher suite configuration and the presence of HSTS headers. A complete SSL check ensures not only that the certificate is valid but that the entire TLS configuration meets current security standards.
Professional uptime monitoring is more than a simple ping test. It is a multi-layered system of synthetic checks, real-user monitoring, escalated alerting chains and transparent customer communication. Investment in professional monitoring pays for itself through faster error detection, shorter outages and the trust that high availability builds with customers and business partners. For an online shop, reachability is not just a technical metric -- it is the fundamental prerequisite for every revenue euro.
Sources and Studies