Monitoring and Logging in Azure Service Fabric

Monitoring and Logging in Service Fabric Applications

Building applications is only half the story — you must monitor and log them properly to ensure smooth, reliable operations. Service Fabric provides built-in tools and APIs to help developers monitor service health and troubleshoot issues.

🛰️ Why Monitoring and Logging are Critical?

• Early detection of issues (failures, slowdowns, bottlenecks).
• Understanding system behavior and usage patterns.
• Essential for troubleshooting and root cause analysis.
• Supports proactive scaling and maintenance.

Real-World Analogy:

Monitoring and logging are like the black box and radar in an aircraft — even if everything looks fine, they continuously record data that’s invaluable during emergencies.

🚀 Monitoring Tools in Service Fabric

1. Service Fabric Explorer (SFX)

A real-time dashboard that shows:

• Cluster health (overall and per node).
• Application health and event logs.
• Replica status and load metrics.

2. Application Health Reports

Services can report their health explicitly using Service Fabric APIs:

var healthInfo = new HealthInformation("MyApp", "DatabaseConnectivity", HealthState.Warning)
{
    Description = "Database connection slow."
};
await this.Partition.ReportPartitionHealthAsync(healthInfo);
    

Key Points: Healthy, Warning, Error states can be reported.

3. Event Tracing for Windows (ETW)

• Built-in support for high-performance event logging.
• View ETW logs using tools like PerfView, Windows Event Viewer, or Azure Monitor.

4. Application Insights Integration

• Can be easily integrated into Service Fabric services.
• Provides dashboards for request rates, response times, failure rates, etc.
• Custom telemetry like traces, exceptions, dependencies.

🛠️ Basic Logging in Service Fabric Services

Using Built-in EventSource Class

Example:

public static class MyServiceEventSource
{
    private static readonly EventSource eventSource = new EventSource("MyServiceEvents");

    public static void ServiceMessage(StatelessService service, string message)
    {
        eventSource.Write("ServiceMessage", new { ServiceName = service.Context.ServiceName, Message = message });
    }
}
    

Calling Logger inside Services

MyServiceEventSource.ServiceMessage(this, "Order processed successfully!");
    

📈 Visual Monitoring Flow

[Service Health Events] 
    → [Service Fabric Explorer Dashboard] 
        → [Alerts, Warnings, Errors Visible]
            → [Root Cause Analysis and Actions]
    

💡 Did You Know?

⚡ Common Beginner Issues and Solutions

• Problem: Missing health reports.
  Solution: Ensure services call health APIs periodically.
• Problem: Logs not visible in Event Viewer.
  Solution: Check if ETW providers are enabled and logs are written correctly.
• Problem: High latency in monitoring dashboard.
  Solution: Reduce the number of frequent health reports; aggregate where possible.

🚨 Best Practices

• Send regular health checks from services (heartbeat).
• Integrate Application Insights early — much easier than retrofitting later.
• Use structured logging (JSON format) when logging complex data.

✅ Self-Check Quiz

• What are the three health states that can be reported?
• Name one tool you can use to view ETW logs.
• Why is Application Insights useful for Service Fabric applications?