ArcadeDB Adds OpenTelemetry Tracing and Kubernetes Health Probes

ArcadeDB has introduced four independently shippable, fully opt-in pillars to close the observability gap for Kubernetes-native deployments: always-on RED-style metrics with OTLP export, OpenTelemetry-based distributed tracing, structured JSON logging with correlation IDs, and a new Kubernetes liveness endpoint. The design reuses Micrometer's Observation API to instrument hot paths—HTTP handling, query and command execution, transaction commits, and Raft replication—once, so each call site can emit either a metric timer alone or a metric plus a trace span, depending on whether a tracer is registered.

Tracing lives entirely in a separate, SPI-loaded module so the OpenTelemetry SDK and OTLP exporter never touch the core server's classpath unless explicitly enabled. Structured logging ties trace and span IDs together with a request ID, either in JSON log lines or as tags on default text logs, linking metrics, traces, and logs into one coherent story. The new /api/v1/health endpoint gives Kubernetes a cheap, database-independent liveness check, while the existing readiness endpoint can optionally become HA-aware, checking Raft membership and catch-up status.

Every new configuration key defaults to off, and no existing endpoint or behavior changes, meaning teams can upgrade immediately and adopt observability features incrementally, one flag at a time—an approach that eases integration with production-grade Prometheus/Grafana and OpenTelemetry observability stacks.