OS services monitoring

A Service Monitoring Policy defines how Dynatrace monitors an OS service based on its state and specified rules. By default, Dynatrace includes:

• Auto-start Windows OS Services policy
• Auto-start Linux OS Services policy

These policies monitor auto-started services that have a failed status.

Understanding Policy Order

The order of policies is important:

• Policies higher in the list are processed first.
• If a higher policy is fulfilled, lower ones will not be applied.
• This allows for selective alerting with minimal policies.

Example:
If you want to monitor all auto-started services (not just those from Microsoft), add a policy with disabled alerting that checks if the service manufacturer is Microsoft.

Steps to Add a Service Monitoring Policy

1. Navigate to OS Services Monitoring
   • Configure the policy at the Host, Host-Group, or Environment level (refer to previous instructions).
   • Select Add policy to define a new monitoring policy.
2. Define Policy Rules
   • System: Select your OS (Windows or Linux).
   • Rule Name: Enter a descriptive name (this appears in the Summary field).
   • Monitor:
     • Enable monitoring using the OS service availability metric (builtin:osservice.availability).
     • If enabled, Dynatrace will send the service status every 10 seconds (via dt.osservice.status).
     • Note: This metric consumes data points (refer to Metrics powered by Grail).
   • Alert: Enable if you want alerting for this policy.
3. OneAgent Version 1.257+ Settings
   • Alert if service is not installed:
     • Enable this option if you want alerts for missing services on the host.
   • Service Status:
     • Define the service states that should trigger an alert.
     • Example (Windows & Linux):
       • running
       • stopped
       • start_pending
       • stop_pending
       • continue_pending
       • pause_pending
       • paused
4. Using Logic Operations for Service Status Monitoring
   • Example Rules:
     • $eq(running): Triggers an alert if the service is running.
     • $not($eq(paused)): Triggers an alert for any state except paused.
     • $or($eq(paused),$eq(running)): Triggers an alert if the service is paused or running.
5. (Optional) Alerting Delay (OneAgent 1.257+)
   • Define the delay (in 10-second cycles) before an alert is generated.
6. Select Services to Monitor
   • Choose which services to monitor based on specific service properties.

Example Use Case: Alert if a Service is Stopped for More Than 1 Minute

1. Add a new policy and name it Critical Service Monitoring.
2. Enable monitoring using builtin:osservice.availability.
3. Enable alerting.
4. Set the service status condition: $eq(stopped).
5. Set an alerting delay of 6 measurement cycles (since 1 cycle = 10s, 6 cycles = 60s).

This ensures an alert is only triggered if the service remains stopped for more than 1 minute.

Once you have added a service monitoring policy, you need to define which services to monitor by selecting OS Service properties or Host metadata.

Step 1: Add a New Rule

1. Go to OS services monitoring (for the Host, Host-Group, or Environment level).
2. Click Add Rule to define the services to monitor.
3. (Optional) Select Rule Scope:
   • OS Service (default) – Monitors specific services based on properties.
   • Host – Monitors hosts based on metadata.

Step 2: Define Rule Scope

If You Select “Host” (OneAgent version 1.277+)

• Use custom metadata to define which hosts to monitor.
• Define matching conditions for host metadata using string expressions:

Examples of Host Metadata Matching (OneAgent 1.310+)

🔹 Note:

• If your metadata contains special characters (e.g., brackets ( )), escape them with ~.
• Example: $eq(my~(amazing~)property) matches my(amazing)property.

If You Select “OS Service”

Choose which services to monitor based on service properties.

Available Matching Properties (Windows & Linux)

Step 3: Define Matching Conditions

Use string expressions to match services.

Examples of Service Matching

Step 4: Combine Logic Operations (Advanced Matching)

You can combine multiple conditions using AND / OR / NOT operators:

Step 5: Save and Apply

1. Review your monitoring rules.
2. Click Save to apply the configuration.
3. Dynatrace will start monitoring the selected services.

If you are using OneAgent version 1.247+ and Dynatrace version 1.247+, you can add custom properties to your service monitoring policy. These properties allow you to define key-value pairs and customize event messages for better alerting and reporting.

Step 1: Add a Custom Property

1. Go to OS services monitoring (at the Host, Host-Group, or Environment level).
2. Select your monitoring policy or create a new policy.
3. Click Add property to add a custom key-value property.
4. Enter:
   • Key – The name of the property (e.g., "Business Impact").
   • Value – The value assigned to the property (e.g., "Critical").
5. (Optional) Add multiple properties if needed.

Step 2: Customize the Event Details Message

1. Locate the Custom message in the Event details section.
2. Enter a custom message to appear in alerts.

   • Example:

     Service {{ServiceName}} on host {{HostName}} has been down for more than 10 minutes.

   • This message dynamically references monitored data using placeholders like:
     • {{ServiceName}} → The name of the OS service
     • {{HostName}} → The hostname of the affected server

Step 3: Save and Apply Changes

1. Review your policy to ensure accuracy.
2. Click Save changes to apply the policy.
3. Dynatrace will now include custom properties and messages in alerts and events.