chore: add SKIP_MIN_VERSION_CHECK env var and database URL docs

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

CLAUDE.md

@@ -51,10 +51,15 @@ After making code changes, you MUST run the appropriate checks and fix all error
 
 `backend/summarized_schema.txt` provides a compact overview of all tables, columns, types, ENUMs, and foreign keys. Use it to quickly understand the data model and relationships. Note: this file is a simplified summary — it omits indexes, constraints details, and other metadata.
 
-For exact table definitions (indexes, constraints, column defaults, etc.), query the database directly:
+For exact table definitions (indexes, constraints, column defaults, etc.), query the database directly.
+
+**IMPORTANT**: The database name varies per worktree/branch. Always read `.env.local` first to get the correct `DATABASE_URL`. The default `windmill` database is only used on the main branch — worktrees use branch-specific database names (e.g., `windmill_add_fast_filter`). Using the wrong database means your queries, inserts, and migrations will target stale or empty data.
 
 ```bash
-psql postgres://postgres:changeme@localhost:5432/windmill
+# Always check .env.local first:
+grep DATABASE_URL .env.local
+# Then use the correct database URL, e.g.:
+psql "postgres://postgres:changeme@127.0.0.1:5432/windmill_add_fast_filter"
 ```
 
 Useful psql commands:

backend/.sqlx/query-308a85372588477b92cacac56017c9f8112529e676dfb4a8851567ec9842957c.json

@@ -0,0 +1,12 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "WITH batch AS (\n                SELECT c.id FROM v2_job_completed c\n                JOIN v2_job j ON c.id = j.id\n                WHERE j.parent_job IS NULL AND c.fast_filter IS NULL AND c.status != 'skipped'\n                    AND j.kind IN ('script', 'flow', 'singlestepflow')\n                LIMIT 50000\n            )\n            UPDATE v2_job_completed SET fast_filter =\n                CASE WHEN v2_job_completed.status = 'success' THEN 1::smallint ELSE 2::smallint END\n            FROM batch WHERE v2_job_completed.id = batch.id",
+  "describe": {
+    "columns": [],
+    "parameters": {
+      "Left": []
+    },
+    "nullable": []
+  },
+  "hash": "308a85372588477b92cacac56017c9f8112529e676dfb4a8851567ec9842957c"
+}

backend/.sqlx/query-36c4e57afcab22f4b6825ccebe47767b8a8fe0a638250f7c7777e5a9f7530e5c.json

@@ -1,31 +0,0 @@
-{
-  "db_name": "PostgreSQL",
-  "query": "INSERT INTO v2_job_completed AS cj\n                    ( workspace_id\n                    , id\n                    , started_at\n                    , duration_ms\n                    , result\n                    , result_columns\n                    , canceled_by\n                    , canceled_reason\n                    , flow_status\n                    , workflow_as_code_status\n                    , memory_peak\n                    , status\n                    , worker\n                    )\n                SELECT q.workspace_id, q.id, started_at, COALESCE($9::bigint, (EXTRACT('epoch' FROM (now())) - EXTRACT('epoch' FROM (COALESCE(started_at, now()))))*1000), $3, $10, $5, $6,\n                        flow_status, workflow_as_code_status,\n                        $8, CASE WHEN $4::BOOL THEN 'canceled'::job_status\n                        WHEN $7::BOOL THEN 'skipped'::job_status\n                        WHEN $2::BOOL THEN 'success'::job_status\n                        ELSE 'failure'::job_status END AS status,\n                        q.worker\n                FROM v2_job_queue q LEFT JOIN v2_job_status USING (id) WHERE q.id = $1\n            ON CONFLICT (id) DO UPDATE SET status = EXCLUDED.status, result = $3 RETURNING duration_ms AS \"duration_ms!\"",
-  "describe": {
-    "columns": [
-      {
-        "ordinal": 0,
-        "name": "duration_ms!",
-        "type_info": "Int8"
-      }
-    ],
-    "parameters": {
-      "Left": [
-        "Uuid",
-        "Bool",
-        "Jsonb",
-        "Bool",
-        "Varchar",
-        "Text",
-        "Bool",
-        "Int4",
-        "Int8",
-        "TextArray"
-      ]
-    },
-    "nullable": [
-      false
-    ]
-  },
-  "hash": "36c4e57afcab22f4b6825ccebe47767b8a8fe0a638250f7c7777e5a9f7530e5c"
-}

backend/.sqlx/query-77f13b7c1e7e488c6268a9ff6def647e47394bd0a61d44fddb1b2040d05b7f17.json

@@ -1,25 +0,0 @@
-{
-  "db_name": "PostgreSQL",
-  "query": "INSERT INTO v2_job_completed AS cj\n                   ( workspace_id\n                   , id\n                   , duration_ms\n                   , result\n                   , canceled_by\n                   , canceled_reason\n                   , flow_status\n                   , status\n                   , worker\n                )\n                SELECT  q.workspace_id\n                   , q.id\n                   , 0\n                   , $4\n                   , $1\n                   , 'cancel all'\n                   , (SELECT flow_status FROM v2_job_status WHERE id = q.id)\n                   , 'canceled'::job_status\n                   , worker\n        FROM v2_job_queue q\n            JOIN v2_job USING (id)\n        WHERE q.id = any($2) AND running = false AND parent_job IS NULL AND q.workspace_id = $3 AND trigger_kind IS DISTINCT FROM 'schedule'\n            FOR UPDATE SKIP LOCKED\n        ON CONFLICT (id) DO NOTHING RETURNING id AS \"id!\"",
-  "describe": {
-    "columns": [
-      {
-        "ordinal": 0,
-        "name": "id!",
-        "type_info": "Uuid"
-      }
-    ],
-    "parameters": {
-      "Left": [
-        "Varchar",
-        "UuidArray",
-        "Text",
-        "Jsonb"
-      ]
-    },
-    "nullable": [
-      false
-    ]
-  },
-  "hash": "77f13b7c1e7e488c6268a9ff6def647e47394bd0a61d44fddb1b2040d05b7f17"
-}

backend/.sqlx/query-96edefb17b81ab9f7c9f41dd3d5a3663c700583c265fe42edecf8d15cf33b0d0.json

@@ -0,0 +1,32 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "INSERT INTO v2_job_completed AS cj\n                    ( workspace_id\n                    , id\n                    , started_at\n                    , duration_ms\n                    , result\n                    , result_columns\n                    , canceled_by\n                    , canceled_reason\n                    , flow_status\n                    , workflow_as_code_status\n                    , memory_peak\n                    , status\n                    , worker\n                    , fast_filter\n                    )\n                SELECT q.workspace_id, q.id, started_at, COALESCE($9::bigint, (EXTRACT('epoch' FROM (now())) - EXTRACT('epoch' FROM (COALESCE(started_at, now()))))*1000), $3, $10, $5, $6,\n                        flow_status, workflow_as_code_status,\n                        $8, CASE WHEN $4::BOOL THEN 'canceled'::job_status\n                        WHEN $7::BOOL THEN 'skipped'::job_status\n                        WHEN $2::BOOL THEN 'success'::job_status\n                        ELSE 'failure'::job_status END AS status,\n                        q.worker,\n                        $11::smallint\n                FROM v2_job_queue q LEFT JOIN v2_job_status USING (id) WHERE q.id = $1\n            ON CONFLICT (id) DO UPDATE SET status = EXCLUDED.status, result = $3, fast_filter = EXCLUDED.fast_filter RETURNING duration_ms AS \"duration_ms!\"",
+  "describe": {
+    "columns": [
+      {
+        "ordinal": 0,
+        "name": "duration_ms!",
+        "type_info": "Int8"
+      }
+    ],
+    "parameters": {
+      "Left": [
+        "Uuid",
+        "Bool",
+        "Jsonb",
+        "Bool",
+        "Varchar",
+        "Text",
+        "Bool",
+        "Int4",
+        "Int8",
+        "TextArray",
+        "Int2"
+      ]
+    },
+    "nullable": [
+      false
+    ]
+  },
+  "hash": "96edefb17b81ab9f7c9f41dd3d5a3663c700583c265fe42edecf8d15cf33b0d0"
+}

backend/.sqlx/query-ad6d4ffdf39ff238f200323134871ab6cc94fe422eaaff8b48e934136cf4cd16.json

@@ -0,0 +1,12 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "INSERT INTO windmill_migrations (name) VALUES ('fast_filter_indexed')",
+  "describe": {
+    "columns": [],
+    "parameters": {
+      "Left": []
+    },
+    "nullable": []
+  },
+  "hash": "ad6d4ffdf39ff238f200323134871ab6cc94fe422eaaff8b48e934136cf4cd16"
+}

backend/.sqlx/query-e8315dace910c9454a79ef15b9ac6057675c0c697cbd9aa4220eb318fc9515eb.json

@@ -0,0 +1,25 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "INSERT INTO v2_job_completed AS cj\n                   ( workspace_id\n                   , id\n                   , duration_ms\n                   , result\n                   , canceled_by\n                   , canceled_reason\n                   , flow_status\n                   , status\n                   , worker\n                   , fast_filter\n                )\n                SELECT  q.workspace_id\n                   , q.id\n                   , 0\n                   , $4\n                   , $1\n                   , 'cancel all'\n                   , (SELECT flow_status FROM v2_job_status WHERE id = q.id)\n                   , 'canceled'::job_status\n                   , worker\n                   , CASE WHEN kind IN ('script', 'flow', 'singlestepflow') THEN 2::smallint ELSE NULL END\n        FROM v2_job_queue q\n            JOIN v2_job USING (id)\n        WHERE q.id = any($2) AND running = false AND parent_job IS NULL AND q.workspace_id = $3 AND trigger_kind IS DISTINCT FROM 'schedule'\n            FOR UPDATE SKIP LOCKED\n        ON CONFLICT (id) DO NOTHING RETURNING id AS \"id!\"",
+  "describe": {
+    "columns": [
+      {
+        "ordinal": 0,
+        "name": "id!",
+        "type_info": "Uuid"
+      }
+    ],
+    "parameters": {
+      "Left": [
+        "Varchar",
+        "UuidArray",
+        "Text",
+        "Jsonb"
+      ]
+    },
+    "nullable": [
+      false
+    ]
+  },
+  "hash": "e8315dace910c9454a79ef15b9ac6057675c0c697cbd9aa4220eb318fc9515eb"
+}

backend/.sqlx/query-ee47ef28bf714c90dba387b49f92c96bcad6046385da58aa753fb9c89a8cbb63.json

@@ -1,6 +1,6 @@
 {
   "db_name": "PostgreSQL",
-  "query": "\n            INSERT INTO v2_job_completed (\n                id, workspace_id, started_at, completed_at, duration_ms, result, deleted,\n                canceled_by, canceled_reason, flow_status, memory_peak, status, worker,\n                workflow_as_code_status, result_columns, retries, extras\n            ) VALUES (\n                $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17\n            )\n            ON CONFLICT (id) DO NOTHING\n            ",
+  "query": "\n            INSERT INTO v2_job_completed (\n                id, workspace_id, started_at, completed_at, duration_ms, result, deleted,\n                canceled_by, canceled_reason, flow_status, memory_peak, status, worker,\n                workflow_as_code_status, result_columns, retries, extras, fast_filter\n            ) VALUES (\n                $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18\n            )\n            ON CONFLICT (id) DO NOTHING\n            ",
   "describe": {
     "columns": [],
     "parameters": {
@@ -33,10 +33,11 @@
         "Jsonb",
         "TextArray",
         "UuidArray",
-        "Jsonb"
+        "Jsonb",
+        "Int2"
       ]
     },
     "nullable": []
   },
-  "hash": "141a428bf866f2e3d4c6eff756a5fb8c6476d6b9b00b9a336b4616d3b69a491e"
+  "hash": "ee47ef28bf714c90dba387b49f92c96bcad6046385da58aa753fb9c89a8cbb63"
 }

backend/.sqlx/query-f614ef418f7124c789ec17f97f0cf7bed89553063512434119523e89abda3c41.json

@@ -0,0 +1,20 @@
+{
+  "db_name": "PostgreSQL",
+  "query": "SELECT EXISTS(SELECT name FROM windmill_migrations WHERE name = 'fast_filter_indexed')",
+  "describe": {
+    "columns": [
+      {
+        "ordinal": 0,
+        "name": "exists",
+        "type_info": "Bool"
+      }
+    ],
+    "parameters": {
+      "Left": []
+    },
+    "nullable": [
+      null
+    ]
+  },
+  "hash": "f614ef418f7124c789ec17f97f0cf7bed89553063512434119523e89abda3c41"
+}

backend/migrations/20260228000000_add_fast_filter_column.down.sql

@@ -0,0 +1 @@
+ALTER TABLE v2_job_completed DROP COLUMN IF EXISTS fast_filter;

backend/migrations/20260228000000_add_fast_filter_column.up.sql

@@ -0,0 +1 @@
+ALTER TABLE v2_job_completed ADD COLUMN IF NOT EXISTS fast_filter SMALLINT;

backend/windmill-api-jobs/src/execution.rs

@@ -99,6 +99,7 @@ pub async fn cancel_jobs(
                    , flow_status
                    , status
                    , worker
+                   , fast_filter
                 )
                 SELECT  q.workspace_id
                    , q.id
@@ -109,6 +110,7 @@ pub async fn cancel_jobs(
                    , (SELECT flow_status FROM v2_job_status WHERE id = q.id)
                    , 'canceled'::job_status
                    , worker
+                   , CASE WHEN kind IN ('script', 'flow', 'singlestepflow') THEN 2::smallint ELSE NULL END
         FROM v2_job_queue q
             JOIN v2_job USING (id)
         WHERE q.id = any($2) AND running = false AND parent_job IS NULL AND q.workspace_id = $3 AND trigger_kind IS DISTINCT FROM 'schedule'

backend/windmill-api-jobs/src/jobs_export.rs

@@ -358,14 +358,27 @@ pub async fn import_completed_jobs(
         .execute(&mut *tx)
         .await?;
 
+        let fast_filter: Option<i16> = if job.parent_job.is_some()
+            || job.status == JobStatus::Skipped
+            || !matches!(
+                job.kind,
+                JobKind::Script | JobKind::Flow | JobKind::SingleStepFlow
+            ) {
+            None
+        } else if job.status == JobStatus::Success {
+            Some(1)
+        } else {
+            Some(2)
+        };
+
         sqlx::query!(
             r#"
             INSERT INTO v2_job_completed (
                 id, workspace_id, started_at, completed_at, duration_ms, result, deleted,
                 canceled_by, canceled_reason, flow_status, memory_peak, status, worker,
-                workflow_as_code_status, result_columns, retries, extras
+                workflow_as_code_status, result_columns, retries, extras, fast_filter
             ) VALUES (
-                $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17
+                $1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18
             )
             ON CONFLICT (id) DO NOTHING
             "#,
@@ -385,7 +398,8 @@ pub async fn import_completed_jobs(
             job.workflow_as_code_status as _,
             job.result_columns as _,
             job.retries as _,
-            job.extras as _
+            job.extras as _,
+            fast_filter
         )
         .execute(&mut *tx)
         .await?;

backend/windmill-api-jobs/src/query.rs

@@ -8,9 +8,12 @@
 
 //! Query builders for filtering job lists (queue and completed).
 
+use std::sync::atomic::Ordering;
+
 use sql_builder::prelude::*;
 use sql_builder::SqlBuilder;
 use windmill_common::utils::{escape_ilike_pattern, paginate_without_limits, Pagination};
+use windmill_common::FAST_FILTER_INDEXED;
 
 use crate::types::{ListCompletedQuery, ListQueueQuery};
 
@@ -491,6 +494,14 @@ pub fn filter_list_completed_query(
     if let Some(fs) = &lq.has_null_parent {
         if *fs {
             sqlb.and_where_is_null("parent_job");
+
+            if FAST_FILTER_INDEXED.load(Ordering::Relaxed) {
+                if lq.success == Some(false) {
+                    sqlb.and_where("fast_filter = 2");
+                } else if lq.success.is_none() && lq.is_skipped == Some(false) {
+                    sqlb.and_where("fast_filter IS NOT NULL");
+                }
+            }
         }
     }
     if let Some(jk) = &lq.job_kinds {

backend/windmill-api/src/live_migrations.rs

@@ -6,8 +6,12 @@
  * LICENSE-AGPL for a copy of the license.
  */
 
+use std::sync::atomic::Ordering;
+
 use sqlx::Postgres;
 use windmill_common::error::Error;
+use windmill_common::min_version::MIN_VERSION_IS_AT_LEAST_1_647;
+use windmill_common::FAST_FILTER_INDEXED;
 
 use crate::db::{CustomMigrator, DB};
 use sqlx::migrate::Migrate;
@@ -18,6 +22,10 @@ pub async fn custom_migrations(migrator: &mut CustomMigrator, db: &DB) -> Result
         tracing::error!("Could not apply flow versioning fix migration: {err:#}");
     }
 
+    if let Err(err) = fast_filter_migration(db).await {
+        tracing::error!("Could not start fast_filter migration: {err:#}");
+    }
+
     Ok(())
 }
 
@@ -67,3 +75,106 @@ async fn fix_flow_versioning_migration(
     }
     Ok(())
 }
+
+async fn fast_filter_migration(db: &DB) -> Result<(), Error> {
+    let has_done_migration = sqlx::query_scalar!(
+        "SELECT EXISTS(SELECT name FROM windmill_migrations WHERE name = 'fast_filter_indexed')",
+    )
+    .fetch_one(db)
+    .await?
+    .unwrap_or(false);
+
+    if has_done_migration {
+        FAST_FILTER_INDEXED.store(true, Ordering::Relaxed);
+        return Ok(());
+    }
+
+    let db = db.clone();
+    tokio::spawn(async move {
+        if let Err(e) = fast_filter_migration_inner(&db).await {
+            tracing::error!("fast_filter background migration failed: {e:#}");
+        }
+    });
+
+    Ok(())
+}
+
+async fn fast_filter_migration_inner(db: &DB) -> Result<(), Error> {
+    // Wait until all workers are at least version 1.647 so new jobs write fast_filter
+    let skip_version_check = std::env::var("SKIP_MIN_VERSION_CHECK")
+        .map(|v| v == "1" || v == "true")
+        .unwrap_or(false);
+    if !skip_version_check {
+        loop {
+            if MIN_VERSION_IS_AT_LEAST_1_647.met().await {
+                break;
+            }
+            tracing::info!("fast_filter migration: waiting for all workers to be >= 1.647");
+            tokio::time::sleep(std::time::Duration::from_secs(30)).await;
+        }
+    } else {
+        tracing::info!(
+            "fast_filter migration: skipping version check (SKIP_MIN_VERSION_CHECK=true)"
+        );
+    }
+
+    // Backfill existing rows in batches (skip skipped jobs — they stay NULL)
+    loop {
+        let rows_updated = sqlx::query_scalar!(
+            "WITH batch AS (
+                SELECT c.id FROM v2_job_completed c
+                JOIN v2_job j ON c.id = j.id
+                WHERE j.parent_job IS NULL AND c.fast_filter IS NULL AND c.status != 'skipped'
+                    AND j.kind IN ('script', 'flow', 'singlestepflow')
+                LIMIT 50000
+            )
+            UPDATE v2_job_completed SET fast_filter =
+                CASE WHEN v2_job_completed.status = 'success' THEN 1::smallint ELSE 2::smallint END
+            FROM batch WHERE v2_job_completed.id = batch.id"
+        )
+        .execute(db)
+        .await
+        .map_err(|e| Error::internal_err(format!("fast_filter backfill failed: {e:#}")))?
+        .rows_affected();
+
+        tracing::info!("fast_filter migration: backfilled {rows_updated} rows");
+
+        if rows_updated == 0 {
+            break;
+        }
+        tokio::time::sleep(std::time::Duration::from_secs(1)).await;
+    }
+
+    // Create partial indexes concurrently (must be outside a transaction)
+    tracing::info!("fast_filter migration: creating indexes concurrently");
+
+    sqlx::query(
+        "CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_v2_job_completed_fast_filter_not_null
+         ON v2_job_completed (workspace_id, completed_at DESC) WHERE fast_filter IS NOT NULL",
+    )
+    .execute(db)
+    .await
+    .map_err(|e| Error::internal_err(format!("fast_filter index creation failed: {e:#}")))?;
+
+    sqlx::query(
+        "CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_v2_job_completed_fast_filter_failure
+         ON v2_job_completed (workspace_id, completed_at DESC) WHERE fast_filter = 2",
+    )
+    .execute(db)
+    .await
+    .map_err(|e| Error::internal_err(format!("fast_filter index creation failed: {e:#}")))?;
+
+    tracing::info!("fast_filter migration: indexes created, marking migration as done");
+
+    sqlx::query!("INSERT INTO windmill_migrations (name) VALUES ('fast_filter_indexed')")
+        .execute(db)
+        .await
+        .map_err(|e| {
+            Error::internal_err(format!("fast_filter migration flag insert failed: {e:#}"))
+        })?;
+
+    FAST_FILTER_INDEXED.store(true, Ordering::Relaxed);
+    tracing::info!("fast_filter migration: complete");
+
+    Ok(())
+}

backend/windmill-common/src/lib.rs

@@ -196,6 +196,7 @@ lazy_static::lazy_static! {
 
     pub static ref BASE_URL: Arc<RwLock<String>> = Arc::new(RwLock::new("".to_string()));
     pub static ref IS_READY: std::sync::atomic::AtomicBool = std::sync::atomic::AtomicBool::new(false);
+    pub static ref FAST_FILTER_INDEXED: AtomicBool = AtomicBool::new(false);
     pub static ref HUB_BASE_URL: Arc<RwLock<String>> = Arc::new(RwLock::new(DEFAULT_HUB_BASE_URL.to_string()));
 
 

backend/windmill-common/src/min_version.rs

@@ -5,6 +5,7 @@ use tokio::sync::RwLock;
 
 // ============ Feature Definitions ============
 
+pub const MIN_VERSION_IS_AT_LEAST_1_647: VC = vc(1, 647, 0, "Fast filter column");
 pub const MIN_VERSION_SUPPORTS_SYNC_JOBS_DEBOUNCING: VC = vc(1, 602, 0, "Sync jobs debouncing");
 pub const MIN_VERSION_SUPPORTS_DEBOUNCING_V2: VC = vc(1, 597, 0, "Debouncing V2");
 pub const MIN_VERSION_IS_AT_LEAST_1_595: VC = vc(1, 595, 0, "Flow status separate table");

backend/windmill-queue/src/jobs.rs

@@ -923,6 +923,19 @@ async fn commit_completed_job<T: Serialize + Send + Sync + ValidableJson>(
         return value;
     }
 
+    let fast_filter: Option<i16> = if completed_job.parent_job.is_some()
+        || skipped
+        || !matches!(
+            completed_job.kind,
+            JobKind::Script | JobKind::Flow | JobKind::SingleStepFlow
+        ) {
+        None
+    } else if canceled_by.is_some() || !success {
+        Some(2)
+    } else {
+        Some(1)
+    };
+
     let duration =  sqlx::query_scalar!(
             "INSERT INTO v2_job_completed AS cj
                     ( workspace_id
@@ -938,6 +951,7 @@ async fn commit_completed_job<T: Serialize + Send + Sync + ValidableJson>(
                     , memory_peak
                     , status
                     , worker
+                    , fast_filter
                     )
                 SELECT q.workspace_id, q.id, started_at, COALESCE($9::bigint, (EXTRACT('epoch' FROM (now())) - EXTRACT('epoch' FROM (COALESCE(started_at, now()))))*1000), $3, $10, $5, $6,
                         flow_status, workflow_as_code_status,
@@ -945,9 +959,10 @@ async fn commit_completed_job<T: Serialize + Send + Sync + ValidableJson>(
                         WHEN $7::BOOL THEN 'skipped'::job_status
                         WHEN $2::BOOL THEN 'success'::job_status
                         ELSE 'failure'::job_status END AS status,
-                        q.worker
+                        q.worker,
+                        $11::smallint
                 FROM v2_job_queue q LEFT JOIN v2_job_status USING (id) WHERE q.id = $1
-            ON CONFLICT (id) DO UPDATE SET status = EXCLUDED.status, result = $3 RETURNING duration_ms AS \"duration_ms!\"",
+            ON CONFLICT (id) DO UPDATE SET status = EXCLUDED.status, result = $3, fast_filter = EXCLUDED.fast_filter RETURNING duration_ms AS \"duration_ms!\"",
             /* $1 */ completed_job.id,
             /* $2 */ success,
             /* $3 */ result as Json<&T>,
@@ -958,6 +973,7 @@ async fn commit_completed_job<T: Serialize + Send + Sync + ValidableJson>(
             /* $8 */ if mem_peak > 0 { Some(mem_peak) } else { None },
             /* $9 */ duration,
             /* $10 */ result_columns as Option<&Vec<String>>,
+            /* $11 */ fast_filter,
         )
         .fetch_optional(&mut *tx)
         .warn_after_seconds(10)