Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
This article describes a supported telemetry approach for Service Fabric clusters by using Azure Monitor Agent (AMA) and data collection rules (DCRs).
Use this guidance if you're onboarding new monitoring, or if you're migrating from older diagnostics configurations such as Windows Azure Diagnostics (WAD), Linux Azure Diagnostics (LAD), or Log Analytics agent (also known as MMA/OMS).
Why this approach
- WAD and LAD are deprecated and retire on March 31, 2026.
- Log Analytics agent (MMA/OMS) is deprecated and unsupported.
- AMA + DCR is the supported ingestion model for guest logs and performance data in Azure Monitor.
For Azure Monitor migration guidance, see Migrate from Azure Diagnostics extension to Azure Monitor Agent.
Recommended rollout
Use a phased rollout to reduce risk.
- Validate AMA + DCR pipeline.
- Add Service Fabric specific signals.
- Cut over and remove deprecated agents.
Phase 1: Validate AMA + DCR pipeline
- Install AMA on cluster VM scale sets or machines.
- Create and associate at least one DCR. Data collection starts only after association.
- Validate that telemetry is flowing into Azure Monitor Logs.
- Windows event logs into the
Eventtable. - Linux syslog into the
Syslogtable. - Performance counters into the
Perftable.
Key references:
- Manage Azure Monitor Agent
- Create and edit data collection rules
- Data collection rule samples
- Monitor data collection operations
Phase 2: Add Service Fabric signals
Map your required Service Fabric signals and update DCR scope incrementally (for example, by node type canary).
- Windows clusters: collect relevant Service Fabric event channels from Windows Event Log.
- Linux clusters: collect syslog facilities and severities used for Service Fabric events.
- Add required OS and process performance counters for infrastructure monitoring.
Use List of Service Fabric events to determine event IDs and channels that are required for alerting and troubleshooting.
Phase 3: Cut over and remove deprecated agents
After required signals are available through AMA + DCR:
- Remove WAD/LAD extensions.
- Remove Log Analytics agent (MMA/OMS) extensions.
- Validate that alerts, dashboards, and queries continue to work.
- Confirm no persistent DCR ingestion, transformation, or delivery errors.
Minimal validation queries
Windows events
Event
| where TimeGenerated > ago(30m)
| where Source has "ServiceFabric" or RenderedDescription has "Service Fabric"
| take 50
Linux syslog
Syslog
| where TimeGenerated > ago(30m)
| where ProcessName has "ServiceFabric" or SyslogMessage has "ServiceFabric"
| take 50
Performance counters
Perf
| where TimeGenerated > ago(30m)
| summarize AvgValue = avg(CounterValue) by ObjectName, CounterName, InstanceName
| top 50 by AvgValue desc
Next steps
- Review platform diagnostics concepts in Monitor Azure Service Fabric.
- Review available counters and signals in Service Fabric monitoring data reference.
- For Linux event flow details, see Service Fabric Linux cluster events in Syslog.