Observability All Agents
This document describes the configuration options available for exposing metrics and other status endpoints.
Configuration Example
yaml
management:
endpoints:
web:
exposure:
include: [ "health", "info", "prometheus" ]
metrics:
distribution:
slo:
http.server.requests: 25,100,250,500,1000,10000
http.client.requests: 25,100,250,500,1000,2000,3000,4000,5000,6000,7000,8000,9000,10000
deidentify: 25,100,250,500,1000,10000
Fields
endpoints.web.exposure.include
- Description: Specifies the management endpoints that should be accessible via the web.
- Type: List of Strings
- Default: Not specified
- Example:yaml
include: [ "health", "info", "prometheus", "metrics" ]
Common Values:
"health"
: Provides health status information."info"
: Displays general application information."prometheus"
: Enables Prometheus metrics endpoint.
metrics.distribution.slo
The slo
field defines Service Level Objectives for various types of operations. Each key specifies a type of operation, and its value is a comma-separated list of response time thresholds (in milliseconds).
- Type: Map (Key: String, Value: List of Integers)
- Examples:yaml
http.server.requests: 50,100,200,400,800 http.client.requests: 100,200,300,500,1000,5000 deidentify: 100,300,600,1200
Common Keys
http.server.requests
All Agents- Available in:
- Description: SLOs for server-side HTTP request response times.
- Example Values:
25,100,250,500,1000,10000
http.client.requests
All Agents- Description: SLOs for client-side HTTP request response times.
- Example Values:
25,100,250,500,1000,2000,3000,4000,5000,10000
deidentify
Clinical Domain AgentResearch Domain Agent- Description: SLOs for deidentify operation times.
- Example Values:
25,100,250,500,1000,10000
fetchResearchMapping
Research Domain Agent- Description: SLOs for fetching mappings from TCA.
- Example Values:
5,10,25,100,250,500,1000,5000,10000
sendBundleToHds
Research Domain Agent- Description: SLOs for sending bundles to research domain HDS.
- Example Values:
5,10,25,100,250,500,1000,5000,10000
Notes
- Management Endpoints: Ensure only necessary endpoints are exposed for security purposes.
- SLO Thresholds: Choose thresholds that reflect realistic performance expectations for each operation type.
- Granularity: Use detailed thresholds for precise performance tracking and monitoring.