fix(hive-metric): blocking OTLP client (no-reactor panic), reject negative counters, allow negative gauges

hive-metric/Cargo.toml

@@ -13,10 +13,14 @@ clap.workspace = true
 # OTEL SDK — only used by this crate, so kept local rather than in workspace.dependencies.
 opentelemetry = "0.32"
 opentelemetry_sdk = { version = "0.32", features = ["metrics"] }
+# Blocking (not async) reqwest client on purpose: the metrics SDK drives the
+# OTLP push from a `PeriodicReader` background thread that has no Tokio runtime,
+# so the async client panics there with "no reactor running". The blocking
+# client sends on that thread directly — and for a fire-and-forget CLI that
+# records one point then `shutdown()`s, a synchronous send is exactly right.
 opentelemetry-otlp = { version = "0.32", default-features = false, features = [
   "metrics",
   "http-json",
-  "reqwest-client",
+  "reqwest-blocking-client",
   "reqwest-rustls",
 ] }
-tokio = { workspace = true, features = ["rt", "macros"] }