From c41e3adcc60ac7172dffbb2b9e868839202c47e2 Mon Sep 17 00:00:00 2001
From: Alexander Petric <alpetric@users.noreply.github.com>
Date: Tue, 31 Mar 2026 01:55:38 -0400
Subject: [PATCH] db health page improvements (#8631)

* db health page improvements

* Update SQLx metadata

---------

Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
---
 ...d6b6792ae5b97ed90b0ee07d17480b2dffe2.json} |  6 +-
 ...8137e4173a34c54344ee522745640d48b8813.json | 32 ++++++
 ...b2e890bc4146bb0419210cd76743cacda8bfa.json | 44 +++++++++
 ...153c43903f929ae5d62fbba12610f89c36d55.json |  2 +-
 ...3c5798e6f7aca1e6e564330766d77038ef0e3.json | 44 ---------
 backend/windmill-api/src/db_health.rs         | 97 +++++++++++++------
 .../instanceSettings/DbHealth.svelte          | 18 ++--
 7 files changed, 160 insertions(+), 83 deletions(-)
 rename backend/.sqlx/{query-87d07998fe8373f5b89eccf6f0528c02e389bf827d935d867430ad3459104dd9.json => query-07770a002a49428c4f956cfc7262d6b6792ae5b97ed90b0ee07d17480b2dffe2.json} (52%)
 create mode 100644 backend/.sqlx/query-2d95191e899d60385b32f36f2e38137e4173a34c54344ee522745640d48b8813.json
 create mode 100644 backend/.sqlx/query-30930bfb0513f1a70194a900011b2e890bc4146bb0419210cd76743cacda8bfa.json
 delete mode 100644 backend/.sqlx/query-bc54ea311580a0525c1f36aaa543c5798e6f7aca1e6e564330766d77038ef0e3.json

diff --git a/backend/.sqlx/query-87d07998fe8373f5b89eccf6f0528c02e389bf827d935d867430ad3459104dd9.json b/backend/.sqlx/query-07770a002a49428c4f956cfc7262d6b6792ae5b97ed90b0ee07d17480b2dffe2.json
similarity index 52%
rename from backend/.sqlx/query-87d07998fe8373f5b89eccf6f0528c02e389bf827d935d867430ad3459104dd9.json
rename to backend/.sqlx/query-07770a002a49428c4f956cfc7262d6b6792ae5b97ed90b0ee07d17480b2dffe2.json
index 24d15601e7..145fdc1229 100644
--- a/backend/.sqlx/query-87d07998fe8373f5b89eccf6f0528c02e389bf827d935d867430ad3459104dd9.json
+++ b/backend/.sqlx/query-07770a002a49428c4f956cfc7262d6b6792ae5b97ed90b0ee07d17480b2dffe2.json
@@ -1,11 +1,11 @@
 {
   "db_name": "PostgreSQL",
-  "query": "SELECT COUNT(*) as cnt FROM pg_stat_activity WHERE state = 'active'",
+  "query": "SELECT setting::bigint as \"max!\" FROM pg_settings WHERE name = 'max_connections'",
   "describe": {
     "columns": [
       {
         "ordinal": 0,
-        "name": "cnt",
+        "name": "max!",
         "type_info": "Int8"
       }
     ],
@@ -16,5 +16,5 @@
       null
     ]
   },
-  "hash": "87d07998fe8373f5b89eccf6f0528c02e389bf827d935d867430ad3459104dd9"
+  "hash": "07770a002a49428c4f956cfc7262d6b6792ae5b97ed90b0ee07d17480b2dffe2"
 }
diff --git a/backend/.sqlx/query-2d95191e899d60385b32f36f2e38137e4173a34c54344ee522745640d48b8813.json b/backend/.sqlx/query-2d95191e899d60385b32f36f2e38137e4173a34c54344ee522745640d48b8813.json
new file mode 100644
index 0000000000..d31420ada0
--- /dev/null
+++ b/backend/.sqlx/query-2d95191e899d60385b32f36f2e38137e4173a34c54344ee522745640d48b8813.json
@@ -0,0 +1,32 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "SELECT\n            COUNT(*) as \"total!\",\n            COUNT(*) FILTER (WHERE state = 'active') as \"active!\",\n            COUNT(*) FILTER (WHERE state = 'idle') as \"idle!\"\n        FROM pg_stat_activity\n        WHERE backend_type = 'client backend'",
+  "describe": {
+    "columns": [
+      {
+        "ordinal": 0,
+        "name": "total!",
+        "type_info": "Int8"
+      },
+      {
+        "ordinal": 1,
+        "name": "active!",
+        "type_info": "Int8"
+      },
+      {
+        "ordinal": 2,
+        "name": "idle!",
+        "type_info": "Int8"
+      }
+    ],
+    "parameters": {
+      "Left": []
+    },
+    "nullable": [
+      null,
+      null,
+      null
+    ]
+  },
+  "hash": "2d95191e899d60385b32f36f2e38137e4173a34c54344ee522745640d48b8813"
+}
diff --git a/backend/.sqlx/query-30930bfb0513f1a70194a900011b2e890bc4146bb0419210cd76743cacda8bfa.json b/backend/.sqlx/query-30930bfb0513f1a70194a900011b2e890bc4146bb0419210cd76743cacda8bfa.json
new file mode 100644
index 0000000000..ca31cbc9ca
--- /dev/null
+++ b/backend/.sqlx/query-30930bfb0513f1a70194a900011b2e890bc4146bb0419210cd76743cacda8bfa.json
@@ -0,0 +1,44 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "SELECT\n            table_name as \"table_name!\",\n            SUM(live_tuples)::bigint as \"live_tuples!\",\n            SUM(dead_tuples)::bigint as \"dead_tuples!\",\n            MAX(last_autovacuum) as \"last_autovacuum\",\n            MAX(last_autoanalyze) as \"last_autoanalyze\"\n        FROM (\n            SELECT\n                CASE\n                    WHEN i.inhparent IS NOT NULL THEN schemaname || '.' || p.relname\n                    ELSE schemaname || '.' || s.relname\n                END as table_name,\n                COALESCE(n_live_tup, 0) as live_tuples,\n                COALESCE(n_dead_tup, 0) as dead_tuples,\n                last_autovacuum,\n                last_autoanalyze\n            FROM pg_stat_user_tables s\n            LEFT JOIN pg_class c ON c.relname = s.relname AND c.relnamespace = (\n                SELECT oid FROM pg_namespace WHERE nspname = s.schemaname\n            )\n            LEFT JOIN pg_inherits i ON i.inhrelid = c.oid\n            LEFT JOIN pg_class p ON p.oid = i.inhparent\n        ) sub\n        GROUP BY table_name\n        ORDER BY SUM(dead_tuples) DESC",
+  "describe": {
+    "columns": [
+      {
+        "ordinal": 0,
+        "name": "table_name!",
+        "type_info": "Text"
+      },
+      {
+        "ordinal": 1,
+        "name": "live_tuples!",
+        "type_info": "Int8"
+      },
+      {
+        "ordinal": 2,
+        "name": "dead_tuples!",
+        "type_info": "Int8"
+      },
+      {
+        "ordinal": 3,
+        "name": "last_autovacuum",
+        "type_info": "Timestamptz"
+      },
+      {
+        "ordinal": 4,
+        "name": "last_autoanalyze",
+        "type_info": "Timestamptz"
+      }
+    ],
+    "parameters": {
+      "Left": []
+    },
+    "nullable": [
+      null,
+      null,
+      null,
+      null,
+      null
+    ]
+  },
+  "hash": "30930bfb0513f1a70194a900011b2e890bc4146bb0419210cd76743cacda8bfa"
+}
diff --git a/backend/.sqlx/query-5a219a2532517869578c4504ff3153c43903f929ae5d62fbba12610f89c36d55.json b/backend/.sqlx/query-5a219a2532517869578c4504ff3153c43903f929ae5d62fbba12610f89c36d55.json
index 36ddb8ab9f..713ccb9dd3 100644
--- a/backend/.sqlx/query-5a219a2532517869578c4504ff3153c43903f929ae5d62fbba12610f89c36d55.json
+++ b/backend/.sqlx/query-5a219a2532517869578c4504ff3153c43903f929ae5d62fbba12610f89c36d55.json
@@ -15,7 +15,7 @@
       ]
     },
     "nullable": [
-      true
+      null
     ]
   },
   "hash": "5a219a2532517869578c4504ff3153c43903f929ae5d62fbba12610f89c36d55"
diff --git a/backend/.sqlx/query-bc54ea311580a0525c1f36aaa543c5798e6f7aca1e6e564330766d77038ef0e3.json b/backend/.sqlx/query-bc54ea311580a0525c1f36aaa543c5798e6f7aca1e6e564330766d77038ef0e3.json
deleted file mode 100644
index c68971b7ba..0000000000
--- a/backend/.sqlx/query-bc54ea311580a0525c1f36aaa543c5798e6f7aca1e6e564330766d77038ef0e3.json
+++ /dev/null
@@ -1,44 +0,0 @@
-{
-  "db_name": "PostgreSQL",
-  "query": "SELECT\n            schemaname || '.' || relname as \"table_name!\",\n            COALESCE(n_live_tup, 0) as \"live_tuples!\",\n            COALESCE(n_dead_tup, 0) as \"dead_tuples!\",\n            last_autovacuum as \"last_autovacuum\",\n            last_autoanalyze as \"last_autoanalyze\"\n        FROM pg_stat_user_tables\n        ORDER BY n_dead_tup DESC\n        LIMIT 15",
-  "describe": {
-    "columns": [
-      {
-        "ordinal": 0,
-        "name": "table_name!",
-        "type_info": "Text"
-      },
-      {
-        "ordinal": 1,
-        "name": "live_tuples!",
-        "type_info": "Int8"
-      },
-      {
-        "ordinal": 2,
-        "name": "dead_tuples!",
-        "type_info": "Int8"
-      },
-      {
-        "ordinal": 3,
-        "name": "last_autovacuum",
-        "type_info": "Timestamptz"
-      },
-      {
-        "ordinal": 4,
-        "name": "last_autoanalyze",
-        "type_info": "Timestamptz"
-      }
-    ],
-    "parameters": {
-      "Left": []
-    },
-    "nullable": [
-      null,
-      null,
-      null,
-      true,
-      true
-    ]
-  },
-  "hash": "bc54ea311580a0525c1f36aaa543c5798e6f7aca1e6e564330766d77038ef0e3"
-}
diff --git a/backend/windmill-api/src/db_health.rs b/backend/windmill-api/src/db_health.rs
index 91c8013950..5b881331c0 100644
--- a/backend/windmill-api/src/db_health.rs
+++ b/backend/windmill-api/src/db_health.rs
@@ -12,8 +12,6 @@ use serde::{Deserialize, Serialize};
 use windmill_common::error::JsonResult;
 
 use crate::db::{ApiAuthed, DB};
-use crate::health::get_pool_stats;
-use crate::health::PoolStats;
 use crate::utils::require_super_admin;
 
 pub fn global_service() -> Router {
@@ -81,8 +79,10 @@ pub struct LargeResultRow {
 
 #[derive(Serialize)]
 pub struct ConnectionPoolInfo {
-    pub pool: PoolStats,
+    pub pg_max_connections: i64,
+    pub pg_total_connections: i64,
     pub pg_active_connections: i64,
+    pub pg_idle_connections: i64,
     pub status: HealthLevel,
     pub message: String,
 }
@@ -307,16 +307,30 @@ async fn fetch_large_results(
 }
 
 async fn fetch_connection_pool(db: &DB) -> windmill_common::error::Result<ConnectionPoolInfo> {
-    let pool = get_pool_stats(db);
+    let max_row = sqlx::query_scalar!(
+        r#"SELECT setting::bigint as "max!" FROM pg_settings WHERE name = 'max_connections'"#
+    )
+    .fetch_one(db)
+    .await?;
 
-    let active_row =
-        sqlx::query!("SELECT COUNT(*) as cnt FROM pg_stat_activity WHERE state = 'active'")
-            .fetch_one(db)
-            .await?;
+    let stats_row = sqlx::query!(
+        r#"SELECT
+            COUNT(*) as "total!",
+            COUNT(*) FILTER (WHERE state = 'active') as "active!",
+            COUNT(*) FILTER (WHERE state = 'idle') as "idle!"
+        FROM pg_stat_activity
+        WHERE backend_type = 'client backend'"#
+    )
+    .fetch_one(db)
+    .await?;
 
-    let pg_active = active_row.cnt.unwrap_or(0);
-    let utilization = if pool.max_connections > 0 {
-        pool.size as f64 / pool.max_connections as f64
+    let pg_max = max_row;
+    let pg_total = stats_row.total;
+    let pg_active = stats_row.active;
+    let pg_idle = stats_row.idle;
+
+    let utilization = if pg_max > 0 {
+        pg_total as f64 / pg_max as f64
     } else {
         0.0
     };
@@ -325,50 +339,75 @@ async fn fetch_connection_pool(db: &DB) -> windmill_common::error::Result<Connec
         (
             HealthLevel::Green,
             format!(
-                "Pool utilization: {:.0}% ({}/{})",
+                "Connection utilization: {:.0}% ({}/{})",
                 utilization * 100.0,
-                pool.size,
-                pool.max_connections
+                pg_total,
+                pg_max
             ),
         )
     } else if utilization < 0.95 {
         (
             HealthLevel::Yellow,
             format!(
-                "Pool utilization is high: {:.0}% ({}/{}). Consider increasing max_connections.",
+                "Connection utilization is high: {:.0}% ({}/{}). Consider increasing max_connections.",
                 utilization * 100.0,
-                pool.size,
-                pool.max_connections
+                pg_total,
+                pg_max
             ),
         )
     } else {
         (
             HealthLevel::Red,
             format!(
-                "Pool near exhaustion: {:.0}% ({}/{}). Increase max_connections urgently.",
+                "Connections near exhaustion: {:.0}% ({}/{}). Increase max_connections urgently.",
                 utilization * 100.0,
-                pool.size,
-                pool.max_connections
+                pg_total,
+                pg_max
             ),
         )
     };
 
-    Ok(ConnectionPoolInfo { pool, pg_active_connections: pg_active, status, message })
+    Ok(ConnectionPoolInfo {
+        pg_max_connections: pg_max,
+        pg_total_connections: pg_total,
+        pg_active_connections: pg_active,
+        pg_idle_connections: pg_idle,
+        status,
+        message,
+    })
 }
 
 async fn fetch_table_maintenance(
     db: &DB,
 ) -> windmill_common::error::Result<Vec<TableMaintenanceInfo>> {
+    // Aggregate partitioned tables (e.g. audit_YYYYMMDD -> audit_partitioned)
+    // while keeping non-partitioned tables as-is
     let rows = sqlx::query!(
         r#"SELECT
-            schemaname || '.' || relname as "table_name!",
-            COALESCE(n_live_tup, 0) as "live_tuples!",
-            COALESCE(n_dead_tup, 0) as "dead_tuples!",
-            last_autovacuum as "last_autovacuum",
-            last_autoanalyze as "last_autoanalyze"
-        FROM pg_stat_user_tables
-        ORDER BY n_dead_tup DESC
-        LIMIT 15"#
+            table_name as "table_name!",
+            SUM(live_tuples)::bigint as "live_tuples!",
+            SUM(dead_tuples)::bigint as "dead_tuples!",
+            MAX(last_autovacuum) as "last_autovacuum",
+            MAX(last_autoanalyze) as "last_autoanalyze"
+        FROM (
+            SELECT
+                CASE
+                    WHEN i.inhparent IS NOT NULL THEN schemaname || '.' || p.relname
+                    ELSE schemaname || '.' || s.relname
+                END as table_name,
+                COALESCE(n_live_tup, 0) as live_tuples,
+                COALESCE(n_dead_tup, 0) as dead_tuples,
+                last_autovacuum,
+                last_autoanalyze
+            FROM pg_stat_user_tables s
+            LEFT JOIN pg_class c ON c.relname = s.relname AND c.relnamespace = (
+                SELECT oid FROM pg_namespace WHERE nspname = s.schemaname
+            )
+            LEFT JOIN pg_inherits i ON i.inhrelid = c.oid
+            LEFT JOIN pg_class p ON p.oid = i.inhparent
+        ) sub
+        GROUP BY table_name
+        ORDER BY SUM(dead_tuples) DESC"#
     )
     .fetch_all(db)
     .await?;
diff --git a/frontend/src/lib/components/instanceSettings/DbHealth.svelte b/frontend/src/lib/components/instanceSettings/DbHealth.svelte
index 1986e78a9c..9614422552 100644
--- a/frontend/src/lib/components/instanceSettings/DbHealth.svelte
+++ b/frontend/src/lib/components/instanceSettings/DbHealth.svelte
@@ -267,7 +267,13 @@
 								<tbody>
 									{#each data.large_results.top_large_results as r}
 										<tr class="border-surface-secondary border-b last:border-0">
-											<td class="text-primary py-1 pr-4 font-mono">{r.id.substring(0, 8)}...</td>
+											<td class="text-primary py-1 pr-4 font-mono"
+												><a
+													href="/run/{r.id}?workspace={r.workspace_id}"
+													class="text-blue-600 hover:underline dark:text-blue-400"
+													>{r.id.substring(0, 8)}...</a
+												></td
+											>
 											<td class="text-secondary py-1 pr-4">{r.workspace_id}</td>
 											<td class="text-secondary py-1 pr-4 font-mono">{r.runnable_path ?? '-'}</td>
 											<td class="text-secondary py-1 pr-4 text-right"
@@ -292,7 +298,7 @@
 				class="flex w-full items-center justify-between p-3 text-left hover:bg-surface-secondary/50"
 				onclick={() => toggleSection('connection_pool')}
 			>
-				<h3 class="text-primary text-sm font-semibold">Connection Pool</h3>
+				<h3 class="text-primary text-sm font-semibold">Database Connections</h3>
 				<div class="flex items-center gap-2">
 					<span
 						class="rounded px-1.5 py-0.5 text-xs font-medium {statusBadge(
@@ -311,12 +317,12 @@
 			{#if expandedSections.connection_pool}
 				<div class="border-surface-secondary flex flex-col gap-1 border-t p-3 text-xs">
 					<p class="text-secondary">
-						Pool size: <strong>{data.connection_pool.pool.size}</strong> / Max:
-						<strong>{data.connection_pool.pool.max_connections}</strong>
-						/ Idle: <strong>{data.connection_pool.pool.idle}</strong>
+						Total connections: <strong>{data.connection_pool.pg_total_connections}</strong> / Max:
+						<strong>{data.connection_pool.pg_max_connections}</strong>
 					</p>
 					<p class="text-secondary">
-						Active PG connections: <strong>{data.connection_pool.pg_active_connections}</strong>
+						Active: <strong>{data.connection_pool.pg_active_connections}</strong>
+						/ Idle: <strong>{data.connection_pool.pg_idle_connections}</strong>
 					</p>
 					<p class="{statusColor(data.connection_pool.status)} mt-1 font-medium">
 						{data.connection_pool.message}