Allow Custom Events and Alerts to Evaluate Conditions Across Multiple Entity Types

Problem Statement

Instana Custom Events currently require users to select a single entity type (host, service, application, etc.) when defining event conditions. Once the entity type is selected, all event conditions and metrics are restricted to that entity.

This limitation prevents users from creating more sophisticated alerting logic that correlates metrics across multiple entity types within a single event definition.

Many monitoring and observability use cases require evaluating conditions from several layers of the stack simultaneously. For example, users may want to combine application performance metrics, service-level metrics, and infrastructure metrics into a single alert to better determine whether an issue is actionable.

Today, this often results in multiple independent alerts that cannot be correlated within a single event definition, increasing alert noise and making it harder to reduce false positives.

Proposed Enhancement

Allow Custom Events (and ideally Smart Alerts) to define conditions using metrics from multiple entity types within a single event.

Possible implementation approaches:

• Allow each condition to independently select its entity type.
• Support mixed conditions across hosts, services, applications, processes, and infrastructure components.
• Support logical operators (AND / OR) across conditions regardless of entity type.
• Maintain existing team and scope permissions when evaluating conditions.

Example of a desired outcome:

IF Application Error Rate > 5% AND Service Latency > 2 seconds AND Host CPU Utilization > 90% THEN Generate Event; Instead of forcing all conditions to come from a single entity type
 

Business Value

Reduce False Positives

- Many alerts become meaningful only when multiple indicators occur together. Correlating conditions across entity types helps distinguish real incidents from isolated metric spikes. 

Improve Alert Precision

- Teams can create more actionable alerts by incorporating both application and infrastructure context into a single evaluation. 

Simplify Alert Management

- Rather than maintaining multiple independent alerts and manually correlating them during investigations, users can define a single event that reflects the actual health criteria of a service. 

Improve Adoption for AppDynamics Migrations

- Organizations migrating from AppDynamics frequently use health rules that combine multiple metrics and conditions to determine application health. Additional flexibility in event construction would help customers recreate existing monitoring strategies within Instana.