Snowflake Observability

Overview

Set up comprehensive observability for Snowflake using built-in ACCOUNT_USAGE views, Snowflake Alerts, and integration with external monitoring systems.

Prerequisites

• Role with access to SNOWFLAKE.ACCOUNT_USAGE (ACCOUNTADMIN or granted)
• Notification integration configured for alerts
• Optional: Prometheus/Grafana or Datadog for external dashboards

Instructions

Step 1: Key Monitoring Queries

-- === QUERY PERFORMANCE ===
-- Average query time by warehouse (last 7 days)
SELECT warehouse_name,
       COUNT(*) AS query_count,
       ROUND(AVG(total_elapsed_time) / 1000, 1) AS avg_seconds,
       ROUND(PERCENTILE_CONT(0.95) WITHIN GROUP (ORDER BY total_elapsed_time) / 1000, 1) AS p95_seconds,
       ROUND(PERCENTILE_CONT(0.99) WITHIN GROUP (ORDER BY total_elapsed_time) / 1000, 1) AS p99_seconds
FROM SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY
WHERE start_time >= DATEADD(days, -7, CURRENT_TIMESTAMP())
  AND execution_status = 'SUCCESS'
  AND query_type = 'SELECT'
GROUP BY warehouse_name
ORDER BY avg_seconds DESC;

-- === ERROR RATE ===
-- Error rate by hour
SELECT DATE_TRUNC('hour', start_time) AS hour,
       COUNT_IF(execution_status = 'SUCCESS') AS success,
       COUNT_IF(execution_status = 'FAIL') AS failures,
       ROUND(COUNT_IF(execution_status = 'FAIL') * 100.0 /
             NULLIF(COUNT(*), 0), 2) AS error_rate_pct
FROM SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY
WHERE start_time >= DATEADD(hours, -24, CURRENT_TIMESTAMP())
GROUP BY hour
ORDER BY hour;

-- === CREDIT CONSUMPTION ===
-- Hourly credit usage
SELECT DATE_TRUNC('hour', start_time) AS hour,
       warehouse_name,
       SUM(credits_used) AS credits
FROM SNOWFLAKE.ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY
WHERE start_time >= DATEADD(hours, -24, CURRENT_TIMESTAMP())
GROUP BY hour, warehouse_name
ORDER BY hour DESC, credits DESC;

-- === STORAGE GROWTH ===
-- Daily storage trend
SELECT usage_date,
       ROUND(storage_bytes / 1e12, 3) AS storage_tb,
       ROUND(stage_bytes / 1e12, 3) AS stage_tb,
       ROUND(failsafe_bytes / 1e12, 3) AS failsafe_tb
FROM SNOWFLAKE.ACCOUNT_USAGE.STORAGE_USAGE
WHERE usage_date >= DATEADD(days, -30, CURRENT_DATE())
ORDER BY usage_date;

Step 2: Built-in Snowflake Alerts

-- Alert: High error rate
CREATE OR REPLACE ALERT high_error_rate_alert
  WAREHOUSE = ANALYTICS_WH
  SCHEDULE = '15 MINUTE'
  IF (EXISTS (
    SELECT 1
    FROM SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY
    WHERE start_time >= DATEADD(minutes, -15, CURRENT_TIMESTAMP())
    GROUP BY ALL
    HAVING COUNT_IF(execution_status = 'FAIL') * 100.0 / COUNT(*) > 5
  ))
  THEN
    CALL SYSTEM$SEND_EMAIL(
      'ops_notifications',
      'oncall@company.com',
      'Snowflake: Error rate > 5%',
      'Query error rate exceeded 5% in the last 15 minutes.'
    );

-- Alert: Warehouse stuck running (no auto-suspend)
CREATE OR REPLACE ALERT warehouse_running_alert
  WAREHOUSE = ANALYTICS_WH
  SCHEDULE = '60 MINUTE'
  IF (EXISTS (
    SELECT 1 FROM INFORMATION_SCHEMA.WAREHOUSES
    WHERE state = 'STARTED'
      AND DATEDIFF('hour', COALESCE(resumed_on, created_on), CURRENT_TIMESTAMP()) > 4
  ))
  THEN
    CALL SYSTEM$SEND_EMAIL(
      'ops_notifications',
      'ops@company.com',
      'Snowflake: Warehouse running > 4 hours',
      'A warehouse has been running for over 4 hours. Check auto-suspend settings.'
    );

-- Alert: Task failures
CREATE OR REPLACE ALERT task_failure_alert
  WAREHOUSE = ANALYTICS_WH
  SCHEDULE = '10 MINUTE'
  IF (EXISTS (
    SELECT 1
    FROM TABLE(INFORMATION_SCHEMA.TASK_HISTORY(
      SCHEDULED_TIME_RANGE_START => DATEADD(minutes, -10, CURRENT_TIMESTAMP())
    ))
    WHERE state = 'FAILED'
  ))
  THEN
    CALL SYSTEM$SEND_EMAIL(
      'ops_notifications',
      'oncall@company.com',
      'Snowflake: Task Failure',
      'One or more Snowflake tasks failed. Check TASK_HISTORY for details.'
    );

-- Resume all alerts
ALTER ALERT high_error_rate_alert RESUME;
ALTER ALERT warehouse_running_alert RESUME;
ALTER ALERT task_failure_alert RESUME;

Step 3: Export Metrics to External Systems

// src/snowflake/metrics-exporter.ts
// Export Snowflake metrics to Prometheus/Datadog

interface SnowflakeMetrics {
  queryCount: number;
  errorRate: number;
  avgLatencyMs: number;
  p95LatencyMs: number;
  creditsUsed: number;
  activeWarehouses: number;
}

async function collectSnowflakeMetrics(
  conn: snowflake.Connection
): Promise<SnowflakeMetrics> {
  const [queryStats] = await query(conn, `
    SELECT
      COUNT(*) AS query_count,
      COUNT_IF(execution_status = 'FAIL') * 100.0 / NULLIF(COUNT(*), 0) AS error_rate,
      AVG(total_elapsed_time) AS avg_latency,
      PERCENTILE_CONT(0.95) WITHIN GROUP (ORDER BY total_elapsed_time) AS p95_latency
    FROM SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY
    WHERE start_time >= DATEADD(minutes, -5, CURRENT_TIMESTAMP())
  `).then(r => r.rows);

  const [creditStats] = await query(conn, `
    SELECT COALESCE(SUM(credits_used), 0) AS credits
    FROM SNOWFLAKE.ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY
    WHERE start_time >= CURRENT_DATE()
  `).then(r => r.rows);

  const [whStats] = await query(conn, `
    SELECT COUNT_IF(state = 'STARTED') AS active
    FROM INFORMATION_SCHEMA.WAREHOUSES
  `).then(r => r.rows);

  return {
    queryCount: queryStats.QUERY_COUNT,
    errorRate: queryStats.ERROR_RATE,
    avgLatencyMs: queryStats.AVG_LATENCY,
    p95LatencyMs: queryStats.P95_LATENCY,
    creditsUsed: creditStats.CREDITS,
    activeWarehouses: whStats.ACTIVE,
  };
}

// Prometheus exposition format
function formatPrometheus(metrics: SnowflakeMetrics): string {
  return [
    `snowflake_queries_total ${metrics.queryCount}`,
    `snowflake_error_rate_percent ${metrics.errorRate}`,
    `snowflake_avg_latency_ms ${metrics.avgLatencyMs}`,
    `snowflake_p95_latency_ms ${metrics.p95LatencyMs}`,
    `snowflake_credits_used_today ${metrics.creditsUsed}`,
    `snowflake_active_warehouses ${metrics.activeWarehouses}`,
  ].join('\n');
}

Step 4: Operational Dashboard Queries

-- Pipeline health dashboard
SELECT
  'Tasks' AS component,
  COUNT_IF(state = 'started') AS running,
  COUNT_IF(state = 'suspended') AS suspended,
  (SELECT COUNT_IF(state = 'FAILED')
   FROM TABLE(INFORMATION_SCHEMA.TASK_HISTORY(
     SCHEDULED_TIME_RANGE_START => DATEADD(hours, -24, CURRENT_TIMESTAMP())
   ))) AS failures_24h
FROM INFORMATION_SCHEMA.TASKS

UNION ALL

SELECT 'Pipes',
  COUNT_IF(is_autoingest_enabled = 'true'), 0,
  0  -- Check PIPE_USAGE_HISTORY for errors
FROM INFORMATION_SCHEMA.PIPES

UNION ALL

SELECT 'Streams',
  COUNT_IF(stale = FALSE),
  COUNT_IF(stale = TRUE), 0
FROM INFORMATION_SCHEMA.STREAMS;

Error Handling

Resources

• Account Usage Views
• QUERYHISTORY
• WAREHOUSEMETERINGHISTORY
• Snowflake Alerts

Next Steps

For incident response, see snowflake-incident-runbook.