Configuration And Health

ShrikeDB uses configuration files plus environment overrides. Keep production configs versioned and keep feature-flag experiments explicit.

Common Ports

Surface Typical local port

gRPC database service

50251

HTTP/REST/MCP

8182

Admin health

9190

Web console

Address supplied by your deployment.

Your config may differ. Trust the active config and process logs over examples.

Health Endpoints

Endpoint Use

/health/live

Process liveness.

/health/ready

Readiness for traffic.

/health/startup

Startup completion.

/health

Combined health response.

/health/vector/namespaces

Vector namespace inventory.

/health/multimodal/alignment

Cross-modality readiness.

Feature Flags

Feature flags are part of ShrikeDB’s safety model. Examples:

  • SHRIKEDB_LAKEHOUSE_ENABLED

  • SHRIKEDB_UNIFIED_QUERY_EXECUTION

  • SHRIKEDB_UNIFIED_RESULT_CACHE

  • SHRIKEDB_PRODUCT_MORPHISM_WRITE_MODE

  • SHRIKEDB_GDS_PATH_WRITE_MODE

Change process-level settings before a restart and verify the resulting behavior with a bounded workload. See Configuration for configuration precedence and Feature Status for prerequisites.

Logs And Traces

When debugging a query, gather:

  • selected database and profile;

  • request ID or result hash;

  • health status;

  • query-fabric trace;

  • graph analytics trace;

  • lakehouse source/mapping status if relevant;

  • frontend bridge error if the request came through the UI.

This makes it much easier to separate planner issues from frontend rendering or stale-process issues.