fix(telemetry): fail the regression gate on a unit change, and report what it gated

compare_to_baseline took the unit from the baseline entry and dropped the current
run's, and nothing compared the two, so a us -> ms change was scored as a numeric
delta: four keys rewritten to the same physical durations reported 99.9%
improvements and the gate exited 0. prom_queries.py says the baseline preserves
the unit "so the comparator can sanity-check unit drift"; it never did. A unit
mismatch now fails and names both units.

The workflow's step summary printed total, regressions and improvements. total is
every key in the report -- the union of baseline and current -- so it was neither
the baseline count nor what was gated, and missing_in_current was computed and
never printed. A run that gated 16 of 20 keys read as a full comparison. The
comparator now reports a real "compared" count and the summary prints it beside
the not-captured count, with a warning when any key was missed. The table also
refused nothing on a truncated report; existence is not readability.

check_regression_bounds told the operator to add max_abs_increase while reading
max_abs_increase_ms / _us, so following the message added a key nothing reads and
the gate kept failing with no explanation. The committed thresholds use only the
suffixed spelling, so the message was the defect. Its three JSON inputs were also
unchecked: a top-level null, list or number parsed and then died on the first
.get, and a string "metrics" survived the placeholder test and reported its own
characters as gated keys -- wrong advice rather than a crash.

Four tests cover these; all four fail against the previous checker.

.github/workflows/telemetry-validation.yml

@@ -5,7 +5,7 @@
 #
 # This is a separate workflow from the main CI. It runs:
 #   - On manual dispatch (workflow_dispatch)
-#   - On pushes to telemetry-related branches
+#   - On any push that touches one of the paths globs below
 #
 # The workflow is intentionally heavyweight (builds rippled, starts Docker
 # services, runs a multi-node cluster) — it validates the full telemetry
@@ -420,22 +420,51 @@ jobs:
               cat "$TIMINGS" >>"$GITHUB_STEP_SUMMARY"
               echo '```' >>"$GITHUB_STEP_SUMMARY"
           elif [ -f "$REGRESSION" ]; then
-              REGR_COUNT=$(jq -e '.summary.regressions' "$REGRESSION") || REGR_COUNT=0
-              IMPR_COUNT=$(jq -e '.summary.improvements' "$REGRESSION") || IMPR_COUNT=0
-              TOTAL=$(jq -e '.summary.total' "$REGRESSION") || TOTAL=0
-              echo "| Stat | Count |" >>"$GITHUB_STEP_SUMMARY"
-              echo "|------|-------|" >>"$GITHUB_STEP_SUMMARY"
-              echo "| Metrics compared | $TOTAL |" >>"$GITHUB_STEP_SUMMARY"
-              echo "| Regressions | $REGR_COUNT |" >>"$GITHUB_STEP_SUMMARY"
-              echo "| Improvements | $IMPR_COUNT |" >>"$GITHUB_STEP_SUMMARY"
-              echo "" >>"$GITHUB_STEP_SUMMARY"
-              if [ "$REGR_COUNT" -gt 0 ]; then
-                  echo "### Regressions" >>"$GITHUB_STEP_SUMMARY"
+              # Existence is not readability: a truncated report satisfies -f,
+              # and the `|| =0` fallbacks below would then render a clean table
+              # of zeros for a run that compared nothing. Same trap the note
+              # above the baseline parse warns about, so check the shape first.
+              SUMMARY_OK=$(jq -r 'if (.summary | type) == "object" then "yes" else "no" end' \
+                  "$REGRESSION" 2>/dev/null) || SUMMARY_OK=no
+              if [ "$SUMMARY_OK" != "yes" ]; then
+                  echo "## Regression Gate: report unreadable" >>"$GITHUB_STEP_SUMMARY"
                   echo "" >>"$GITHUB_STEP_SUMMARY"
-                  echo "| Metric | Baseline | Current | Δ | % | Unit |" >>"$GITHUB_STEP_SUMMARY"
-                  echo "|--------|---------:|--------:|--:|--:|------|" >>"$GITHUB_STEP_SUMMARY"
-                  jq -r '.metrics[] | select(.regressed) | "| \(.key) | \(.baseline) | \(.current) | \(.delta) | \(.pct_change)% | \(.unit) |"' \
-                      "$REGRESSION" >>"$GITHUB_STEP_SUMMARY"
+                  echo "\`$REGRESSION\` exists but carries no \`summary\` object, so no" \
+                      "table can be rendered. The pass/fail above still comes from the" \
+                      "comparator's exit code." >>"$GITHUB_STEP_SUMMARY"
+                  echo "::error::Regression report is present but has no summary object"
+              else
+                  # No `jq -e`: it exits non-zero when a field is legitimately 0 or
+                  # false, which the fallbacks would silently turn into 0 as well.
+                  REGR_COUNT=$(jq -r '.summary.regressions // 0' "$REGRESSION")
+                  IMPR_COUNT=$(jq -r '.summary.improvements // 0' "$REGRESSION")
+                  TOTAL=$(jq -r '.summary.total // 0' "$REGRESSION")
+                  MISSING_COUNT=$(jq -r '.summary.missing_in_current // 0' "$REGRESSION")
+                  COMPARED=$(jq -r '.summary.compared // 0' "$REGRESSION")
+                  # `total` is every key in the report, i.e. the union of the
+                  # baseline and this run, so it is not what was gated. The
+                  # comparator reports `compared` for that; do not derive it from
+                  # total minus missing, because a key can also be skipped for
+                  # being new or for having no data on either side.
+                  echo "| Stat | Count |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "|------|-------|" >>"$GITHUB_STEP_SUMMARY"
+                  echo "| Metrics in report | $TOTAL |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "| Metrics compared | $COMPARED |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "| Not captured this run | $MISSING_COUNT |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "| Regressions | $REGR_COUNT |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "| Improvements | $IMPR_COUNT |" >>"$GITHUB_STEP_SUMMARY"
+                  echo "" >>"$GITHUB_STEP_SUMMARY"
+                  if [ "$MISSING_COUNT" -gt 0 ]; then
+                      echo "::warning::$MISSING_COUNT baseline metric(s) were not captured this run, so they were not gated"
+                  fi
+                  if [ "$REGR_COUNT" -gt 0 ]; then
+                      echo "### Regressions" >>"$GITHUB_STEP_SUMMARY"
+                      echo "" >>"$GITHUB_STEP_SUMMARY"
+                      echo "| Metric | Baseline | Current | Δ | % | Unit |" >>"$GITHUB_STEP_SUMMARY"
+                      echo "|--------|---------:|--------:|--:|--:|------|" >>"$GITHUB_STEP_SUMMARY"
+                      jq -r '.metrics[] | select(.regressed) | "| \(.key) | \(.baseline) | \(.current) | \(.delta) | \(.pct_change)% | \(.unit) |"' \
+                          "$REGRESSION" >>"$GITHUB_STEP_SUMMARY"
+                  fi
               fi
           fi