obs(grafana): PR787 r1 — coderabbit critical (ID collision) + Major (missing node_id filter)

Two findings on PR #787 Lua VM Pool dashboard panels:

## coderabbit Critical (line 1749)

The new "Lua VM Pool" row used `"id": 24`, which collided
with an existing panel in the "Hot Path (legacy PR #560)" row
(line 1234) that already had `"id": 24`. Grafana requires
unique IDs across all rows and panels within a dashboard;
the collision would cause dashboard import failures or
unexpected UI behaviour on render.

Fix: changed the new row's id from `24` to `33`. The
existing dashboard ID space went up through `32` on the
last panel of the previous row, so `33` is the next free
slot and respects the existing numbering convention.

## coderabbit Major (line 1817)

The new Lua VM Pool queries hardcoded `job="elastickv"`
without the `node_id=~"$node_id"` template-variable filter
that every other row in this dashboard uses. Operators
selecting a specific node in the $node_id dropdown expect
all panels to filter — without the filter, the Lua VM Pool
row would always show metrics for every node regardless of
the selection, preventing per-node drilldown when
troubleshooting pool sizing.

Fix: added `node_id=~"$node_id"` to all seven query
expressions in the row:

  - line 1817 — `elastickv_lua_pool_idle` (Idle vs cap)
  - line 1825 — `elastickv_lua_pool_max_idle` (Idle vs cap)
  - line 1890 — `sum by (node_id) (rate(elastickv_lua_pool_hits_total ...))`
  - line 1898 — `sum by (node_id) (rate(elastickv_lua_pool_misses_total ...))`
  - line 1906 — `sum by (node_id) (rate(elastickv_lua_pool_drops_total ...))`
  - line 1970 — `sum by (node_id) (increase(... pool_drops ...))`
  - line 2036 — saturation ratio (idle / max_idle)

The filter pattern (`job="elastickv",node_id=~"$node_id"`)
matches the existing Hot Path row's queries exactly.

Verified:
  - `jq -e . monitoring/grafana/dashboards/elastickv-redis-summary.json` → valid JSON
  - `grep elastickv_lua_pool ... | grep -v node_id` returns only the
    description string (no remaining queries without the filter)

JSON-only change. No Go code touched; no caller audit needed.

Remaining gemini medium (line 86 collector-registration noisy
logs) is below the loop's P0/P1/high/critical/major threshold
and is deferred.

Author: bootjp

monitoring/grafana/dashboards/elastickv-redis-summary.json

@@ -1746,7 +1746,7 @@
         "x": 0,
         "y": 71
       },
-      "id": 24,
+      "id": 33,
       "panels": [
         {
           "datasource": "$datasource",
@@ -1814,15 +1814,15 @@
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "elastickv_lua_pool_idle{job=\"elastickv\"}",
+              "expr": "elastickv_lua_pool_idle{job=\"elastickv\",node_id=~\"$node_id\"}",
               "legendFormat": "idle {{node_id}}",
               "range": true,
               "refId": "A"
             },
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "elastickv_lua_pool_max_idle{job=\"elastickv\"}",
+              "expr": "elastickv_lua_pool_max_idle{job=\"elastickv\",node_id=~\"$node_id\"}",
               "legendFormat": "max_idle {{node_id}}",
               "range": true,
               "refId": "B"
@@ -1887,23 +1887,23 @@
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "sum by (node_id) (rate(elastickv_lua_pool_hits_total{job=\"elastickv\"}[5m]))",
+              "expr": "sum by (node_id) (rate(elastickv_lua_pool_hits_total{job=\"elastickv\",node_id=~\"$node_id\"}[5m]))",
               "legendFormat": "hits {{node_id}}",
               "range": true,
               "refId": "A"
             },
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "sum by (node_id) (rate(elastickv_lua_pool_misses_total{job=\"elastickv\"}[5m]))",
+              "expr": "sum by (node_id) (rate(elastickv_lua_pool_misses_total{job=\"elastickv\",node_id=~\"$node_id\"}[5m]))",
               "legendFormat": "misses {{node_id}}",
               "range": true,
               "refId": "B"
             },
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "sum by (node_id) (rate(elastickv_lua_pool_drops_total{job=\"elastickv\"}[5m]))",
+              "expr": "sum by (node_id) (rate(elastickv_lua_pool_drops_total{job=\"elastickv\",node_id=~\"$node_id\"}[5m]))",
               "legendFormat": "drops {{node_id}}",
               "range": true,
               "refId": "C"
@@ -1967,7 +1967,7 @@
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "sum by (node_id) (increase(elastickv_lua_pool_drops_total{job=\"elastickv\"}[$__range]))",
+              "expr": "sum by (node_id) (increase(elastickv_lua_pool_drops_total{job=\"elastickv\",node_id=~\"$node_id\"}[$__range]))",
               "legendFormat": "{{node_id}}",
               "range": true,
               "refId": "A"
@@ -2033,7 +2033,7 @@
             {
               "datasource": "$datasource",
               "editorMode": "code",
-              "expr": "elastickv_lua_pool_idle{job=\"elastickv\"} / elastickv_lua_pool_max_idle{job=\"elastickv\"}",
+              "expr": "elastickv_lua_pool_idle{job=\"elastickv\",node_id=~\"$node_id\"} / elastickv_lua_pool_max_idle{job=\"elastickv\",node_id=~\"$node_id\"}",
               "legendFormat": "{{node_id}}",
               "range": true,
               "refId": "A"