vs-health

Shared subsystem health tracking for Craton Shield runtimes.

Overview

This no_std, heap-free crate centralizes the per-subsystem health bookkeeping used by the Craton Shield runtime crates. It fixes a small family of bugs previously found in hand-rolled per-runtime trackers: ts_us == 0 auto-recovery deadlocks, shutdown paths that left degraded_since_us populated, and health-attribution drift between status fields and the alert log. All status mutation flows through a single with_subsystem_alert entry point that pairs status changes with a clamped (non-zero) alert timestamp and a dirty bit.

Quick Start

use vs_health::{HealthRegistry, SubsystemId};
use vs_types::AlertSeverity;

let mut registry = HealthRegistry::new();
registry.mark_initialized(SubsystemId::Can);

registry.with_subsystem_alert(
    SubsystemId::Can,
    1_000, // ts_us (monotonic microseconds)
    AlertSeverity::Medium,
    |handle| {
        handle.mark_degraded();
    },
);

assert!(registry.is_dirty());

SubsystemId Variants

Can, Eth, V2x, Diag, SignalIds, Mqtt, CoAp, Ble, Zigbee, LoRa, ModbusEmb, ModbusInd, OpcUa, Profinet, EthernetIp, Dnp3, BacNet, S7Comm, Iec60870, Iec61850, Mms, Goose.

SubsystemId::COUNT and SubsystemId::ALL are exposed for fixed-size lookup tables; the enum is #[non_exhaustive].

License

Apache-2.0. See LICENSE. For workspace overview see ../../README.md.