Module runtime_diagnostics 

Stable Tokio runtime diagnostics (no tokio_unstable required).

Implements step 1 of the fast-Tokio principles: measure first. We capture the stable subset of tokio::runtime::RuntimeMetrics — worker count, alive tasks, global (injection) queue depth, and total worker busy time. The article’s headline signal is the global queue staying deep (“in a healthy application it should generally stay close to empty”); that metric is stable, so it ships here.

Per-worker local-queue depth, steal/overflow counts, blocking-pool depth, and the poll-time/schedule-latency histograms are all gated behind RUSTFLAGS="--cfg tokio_unstable" in current Tokio. Enabling that cfg changes the whole dependency graph build, so it stays opt-in follow-up work rather than a default here.

Structs

RuntimeSnapshot

Snapshot of stable runtime counters at one instant.

Functions

configured_worker_threads

Resolve an explicit worker-thread count for the main runtime.

log_snapshot

Log one snapshot at DEBUG level; no-op unless diagnostics are enabled.

runtime_diagnostics_enabled

Returns true when runtime diagnostics logging is enabled.

snapshot

Capture stable counters from a runtime handle.

spawn_periodic_reporter

Spawn a low-frequency reporter for long-lived processes.