HyperFleet Sentinel Release v0.2.0 includes significant enhancements to observability, decision engine capabilities, and API compatibility.
Container Image
quay.io/openshift-hyperfleet/hyperfleet-sentinel:v0.2.0
Deployment
Deploy via the hyperfleet-infra umbrella Helm chart.
Major Features
- OpenTelemetry Tracing (#51, #73, #79):
- Complete tracing implementation with gRPC/HTTP exporters
- Support for all OTEL sampler types with configurable sampling
- Trace context propagation to CloudEvents via traceparent extension
- Span instrumentation for API calls, event publishing, and polling cycles
- Protocol selection via OTEL_EXPORTER_OTLP_PROTOCOL
- Enabled by default with comprehensive documentation
- CEL-based Decision Engine (#72): Replaced hardcoded max_age with configurable CEL expression evaluation
- Mock HyperFleet API (#76): Development API for testing with production-matching payload generation
- Selective Query Filtering (#60): Condition-based filtering for resource queries
- Poll Staleness Detection (#70): Dead man's switch gauge metric with configurable alert thresholds
- HyperShift HostedCluster Spec (#80): Migrated from OCM spec to HyperShift HostedCluster specification
Fixes
- Fix sentinel delays first event publish for newly created clusters (#77)
- Add fail fast check on HyperFleet API endpoint (#74)
- Add structured logging to health check failure paths (#65)
Helm Chart Improvements
- Align chart with Helm conventions standard (#78)
- Use image.tag for app.kubernetes.io/version label (#81)
- Use 0.0.0-dev version for dev image builds (#69)