Skip to content

[Feature] Translate exception from application insight into incident in SRE #185

@ehsaneskandariboc

Description

@ehsaneskandariboc

Feature description

Enable native, direct integration between Azure Application Insights exceptions and Azure SRE Agent, exceptions can be automatically ingested and translated into incidents without requiring an intermediate Azure Monitor Alert.
This capability would allow the SRE Agent to directly consume exception telemetry (e.g., failures, unhandled exceptions, critical errors) from Application Insights and create actionable incidents in near real time.

When there is a new exception in my app from production, I should see that exception as an incident in Azure SRE agent with azure monitor integration

Use case

Today, when an exception occurs in a production application (e.g., web app or API), it is captured in Application Insights but does not automatically surface as an incident in Azure SRE Agent.
To operationalize this, teams must first define Azure Monitor alerts based on logs or metrics, which introduces:

Additional configuration overhead
Alert latency
Risk of misconfiguration (missed or noisy alerts)

This feature would simplify and accelerate incident detection by allowing:

Immediate visibility of production exceptions as incidents in SRE Agent
Reduced mean time to detect (MTTD)
Better alignment between application telemetry and incident management

This is especially valuable for cloud-native Azure workloads (App Service, AKS, Container Apps, Functions, APIs) where high-volume telemetry already exists in Application Insights.

Direct integration between app insight exceptions and SRE agent instead of having alert in the middle

Current workaround

Currently, teams must:

Capture exceptions in Application Insights
Create Azure Monitor Log or Metric Alerts (e.g., KQL-based queries)
Configure alert rules with thresholds and action groups
Integrate those alerts with Azure SRE Agent

Limitations of this approach:

Adds operational complexity and maintenance overhead
Introduces delay between exception occurrence and incident creation
Requires tuning thresholds to avoid alert fatigue
Creates duplication between telemetry and alert definitions

Proposed approach

Introduce a native connector or ingestion path between Application Insights and Azure SRE Agent with the following capabilities:

Direct ingestion pipeline

Allow SRE Agent to subscribe to exception telemetry from Application Insights (e.g., via diagnostic pipeline or event stream)

Built-in exception-to-incident mapping

Automatically translate exceptions into incidents based on:

Severity level (e.g., Error, Critical)
Exception type or failure count
Impact signals (e.g., request failure rate)

Smart filtering & deduplication

Group similar exceptions into a single incident
Apply noise reduction using AI/ML (e.g., anomaly detection, pattern grouping)

Configurable policies (without alerts)

Define rules directly in SRE Agent such as:

“Create incident when new exception type appears”
“Trigger incident if exception count exceeds X within Y minutes”

Deep linking & context

Include direct links to:

Application Insights traces
Transaction search
Distributed tracing view

End-to-end integration

Seamless experience with Azure Monitor and SRE Agent without forcing users to create intermediate alert rules

This would streamline observability-to-incident workflows and significantly reduce the operational friction for teams running production-grade Azure applications.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions