* feat(#105): add fault correlation engine for root-cause analysis Implement hierarchical and auto-cluster correlation modes to reduce fault noise by identifying root causes and grouping related faults. Key features: - Hierarchical mode: detect root cause → symptom relationships - Auto-cluster mode: group similar faults within time window - YAML-based configuration with pattern wildcards - Muted faults tracking and auto-clear on root cause resolution - Gateway API support with include_muted/include_clusters params New messages: MutedFaultInfo.msg, ClusterInfo.msg Extended: GetFaults.srv, ClearFault.srv, FaultEvent.msg * docs(#105): add correlation endpoints to Postman collection Add new requests demonstrating fault correlation query parameters: - GET Faults with Correlation Data (include_muted + include_clusters) - GET Faults with Muted Details Only - GET Faults with Cluster Details Only Update existing endpoint descriptions to document: - muted_count/cluster_count fields in GET /faults response - auto_cleared_codes in DELETE clear fault response - Faults folder description with correlation feature overview * docs(#105): add fault correlation tutorial Add comprehensive tutorial covering: - Hierarchical mode (root cause → symptom relationships) - Auto-cluster mode (grouping similar faults) - YAML configuration with patterns and rules - REST API query parameters for correlation data - Complete example configuration - Troubleshooting guide * style: fix clang-format and flake8 formatting issues * fix(#105): address thread safety and retroactive cluster muting - Fix data race in PatternMatcher::get_compiled() by adding mutex protection for the mutable compiled_cache_ member - Add retroactive_mute_codes field to ProcessFaultResult for tracking faults that should have been muted before cluster activation - Populate retroactive_mute_codes when cluster transitions from pending to active (reaches min_count threshold) - Add debug logging for retroactive muting in fault_manager_node - Add unit test AutoClusterRetroactiveMuting to verify behavior * fix(#105): add periodic cleanup timer for correlation engine Add 5-second timer to periodically call cleanup_expired() on the correlation engine. This prevents stale pending root causes and pending clusters from accumulating in memory when faults are reported infrequently. Add unit tests: - CleanupExpiredRemovesPendingRootCauses - CleanupExpiredRemovesPendingClusters * perf(#105): minor optimizations and documentation - Skip regex compilation for non-wildcard patterns (perf improvement) - Add validation error for min_count=0 in auto-cluster rules - Add TODO documenting known limitation: pending_clusters_ not cleaned up when fault is cleared before cluster reaches min_count * fix(#105): address code review feedback for fault correlation - Add inline symptom codes support in YAML config (codes: [...] syntax) - Refactor severity_to_string/severity_rank to shared types.hpp/cpp - Add correlation.cleanup_interval_sec parameter (default 5.0s) - Update TODO(#105) → TODO(#127) for pending_clusters cleanup - Add PlantUML sequence diagrams to fault-correlation tutorial Addresses review comments: - Implement missing inline codes parsing (#1) - Fix orphaned TODO reference (#2) - Consolidate duplicated severity functions (#3) - Parameterize magic number 5s (#4) - Add flow diagrams to docs (#6)

Add basic folder structure #3

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions