Merge remote-tracking branch 'origin/staging' into worktree-dynatrace-unmute-check

# Conflicts:
#	apps/sim/tools/generated/tool-metadata.ts

Author: waleedlatif1

apps/docs/content/docs/en/integrations/snowflake.mdx

@@ -164,7 +164,7 @@ Insert structured JSON rows using bound values.
 | `database` | string | Yes | Database name |
 | `schema` | string | Yes | Schema name |
 | `table` | string | Yes | Target Snowflake table name within the selected database and schema context |
-| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest. |
+| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead. |
 
 #### Output
 
@@ -210,7 +210,7 @@ Update matching rows with a bound MERGE statement without inserting new rows.
 | `database` | string | Yes | Database name |
 | `schema` | string | Yes | Schema name |
 | `table` | string | Yes | Target Snowflake table name within the selected database and schema context |
-| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest. |
+| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead. |
 | `matchColumns` | array | Yes | Columns used to match target rows. Match values must be non-null and unique across the submitted rows. |
 
 #### Output
@@ -257,7 +257,7 @@ Update matching rows and insert unmatched rows with a bound MERGE statement.
 | `database` | string | Yes | Database name |
 | `schema` | string | Yes | Schema name |
 | `table` | string | Yes | Target Snowflake table name within the selected database and schema context |
-| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest. |
+| `rows` | json | Yes | Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead. |
 | `matchColumns` | array | Yes | Columns used to match target rows. Match values must be non-null and unique across the submitted rows. |
 
 #### Output

apps/sim/blocks/blocks/snowflake.ts

@@ -649,7 +649,7 @@ export const SnowflakeBlock: BlockConfig<SnowflakeStatementResponse> = {
     rows: {
       type: 'string',
       description:
-        'Structured rows as a JSON array. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest.',
+        'Structured rows as a JSON array. For bulk loads, stage the files and use Load Data instead.',
     },
     matchColumns: {
       type: 'string',

apps/sim/tools/generated/tool-metadata.ts

@@ -81,7 +81,7 @@ export const insertRowsTool: ToolConfig<SnowflakeInsertRowsParams, SnowflakeStat
       required: true,
       visibility: 'user-or-llm',
       description:
-        'Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest.',
+        'Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead.',
     },
   },
   request: snowflakeStatementRequest((params) =>

apps/sim/tools/snowflake/insert_rows.ts

@@ -172,45 +172,6 @@ describe('Snowflake SQL builders', () => {
     )
   })
 
-  it('caps the number of rows per structured write', () => {
-    const rows = Array.from({ length: 1001 }, (_, index) => ({ id: index }))
-    expect(() => buildInsertRows({ ...table, rows })).toThrow('cannot exceed 1000 per call')
-    expect(() => buildInsertRows({ ...table, rows })).toThrow('snowflake_load_data')
-    expect(() => buildInsertRows({ ...table, rows: [{ blob: 'x'.repeat(1_000_001) }] })).toThrow(
-      'statement budget'
-    )
-  })
-
-  it('counts the bound value budget in UTF-8 bytes, not UTF-16 code units', () => {
-    expect(() => buildInsertRows({ ...table, rows: [{ blob: '中'.repeat(999_999) }] })).toThrow(
-      'statement budget'
-    )
-    expect(() =>
-      buildInsertRows({ ...table, rows: [{ blob: '中'.repeat(333_333) }] })
-    ).not.toThrow()
-  })
-
-  it('applies the same statement budget to explicit bindings', () => {
-    expect(() =>
-      normalizeBindings({ '1': { type: 'TEXT', value: 'x'.repeat(1_000_001) } })
-    ).toThrow('statement budget')
-    expect(() =>
-      normalizeBindings({
-        '1': { type: 'TEXT', value: 'x'.repeat(600_000) },
-        '2': { type: 'TEXT', value: 'x'.repeat(600_000) },
-      })
-    ).toThrow('statement budget')
-    expect(() =>
-      buildCallProcedure({
-        ...context,
-        database: 'ANALYTICS',
-        schema: 'PUBLIC',
-        procedureName: 'REFRESH_MODEL',
-        procedureArguments: [{ type: 'TEXT', value: '中'.repeat(999_999) }],
-      })
-    ).toThrow('statement budget')
-  })
-
   it('routes a mixed semi-structured column through a single whole-column PARSE_JSON', () => {
     const result = buildInsertRows({
       ...table,

apps/sim/tools/snowflake/sql.test.ts

@@ -90,26 +90,6 @@ function requireQueryId(queryId: string): string {
   return trimmed
 }
 
-/**
- * Snowflake recommends limiting query text to 1 MB per statement, and that limit explicitly covers
- * values supplied through bindings. Statements above it still execute but are truncated before the
- * metadata store persists them, so they can no longer be retried or inspected.
- *
- * The budget is a byte budget, so it is measured against the UTF-8 encoding rather than the
- * JavaScript string length — a multi-byte string is up to 3x longer on the wire than in code units.
- */
-const MAX_BOUND_VALUE_BYTES = 1_000_000
-
-const BULK_INGEST_HINT = 'stage the data and use snowflake_load_data for bulk ingest'
-
-function assertBoundBytesWithinBudget(boundBytes: number): void {
-  if (boundBytes > MAX_BOUND_VALUE_BYTES) {
-    throw new Error(
-      `Snowflake bound values exceed the ${MAX_BOUND_VALUE_BYTES} byte statement budget; send fewer rows per call or ${BULK_INGEST_HINT}`
-    )
-  }
-}
-
 export function normalizeBindings(
   input?: Record<string, SnowflakeBinding>
 ): Record<string, SnowflakeBinding> | undefined {
@@ -119,7 +99,6 @@ export function normalizeBindings(
   }
   const normalized: Record<string, SnowflakeBinding> = {}
   let hasBindings = false
-  let boundBytes = 0
   for (const position in input) {
     if (!Object.hasOwn(input, position)) continue
     hasBindings = true
@@ -136,27 +115,16 @@ export function normalizeBindings(
     if (typeof binding.value !== 'string') {
       throw new Error(`binding ${position} value must be a string`)
     }
-    boundBytes += Buffer.byteLength(binding.value, 'utf8')
-    assertBoundBytesWithinBudget(boundBytes)
     normalized[position] = { type: binding.type, value: binding.value }
   }
   return hasBindings ? normalized : undefined
 }
 
-/**
- * Structured writes build a single statement holding every row, so the row count is capped well
- * below the byte budget to keep a typical write far away from the 1 MB statement recommendation.
- */
-const MAX_WRITE_ROWS = 1_000
-
 class BindingsBuilder {
   readonly bindings: Record<string, SnowflakeBinding> = {}
   private position = 0
-  private boundBytes = 0
 
   private addBinding(type: SnowflakeBinding['type'], value: string): string {
-    this.boundBytes += Buffer.byteLength(value, 'utf8')
-    assertBoundBytesWithinBudget(this.boundBytes)
     this.position += 1
     const key = String(this.position)
     this.bindings[key] = { type, value }
@@ -195,11 +163,6 @@ class BindingsBuilder {
 
 function validateRows(rows: Array<Record<string, unknown>>): string[] {
   if (!Array.isArray(rows) || rows.length === 0) throw new Error('rows must be a non-empty array')
-  if (rows.length > MAX_WRITE_ROWS) {
-    throw new Error(
-      `rows cannot exceed ${MAX_WRITE_ROWS} per call; send the rows in smaller batches or ${BULK_INGEST_HINT}`
-    )
-  }
   const columns = Object.keys(rows[0] ?? {})
   if (columns.length === 0) throw new Error('rows must contain at least one column')
   const identifierKeys = new Set<string>()

apps/sim/tools/snowflake/sql.ts

@@ -91,7 +91,7 @@ export const updateRowsTool: ToolConfig<SnowflakeUpdateRowsParams, SnowflakeStat
       required: true,
       visibility: 'user-or-llm',
       description:
-        'Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest.',
+        'Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead.',
     },
     matchColumns: {
       type: 'array',

apps/sim/tools/snowflake/update_rows.ts

@@ -91,7 +91,7 @@ export const upsertRowsTool: ToolConfig<SnowflakeUpsertRowsParams, SnowflakeStat
       required: true,
       visibility: 'user-or-llm',
       description:
-        'Non-empty JSON array of row objects with matching keys. Max 1000 rows and 1 MB of bound data per call - stage the files and use Load Data for bulk ingest.',
+        'Non-empty JSON array of row objects with matching keys. For bulk loads, stage the files and use Load Data instead.',
     },
     matchColumns: {
       type: 'array',