SQL Server on Google Cloud SQL metrics and alerts
Published 02 July 2026
This documentation contains a list of metrics and alerts that are available when monitoring SQL Server instances on Google Cloud SQL, and the requirements for collecting them. If you can't find a specific metric or alert you would like to use when monitoring your RDS instances, send us feedback.
Metrics
Machine metrics
Because Google Cloud SQL is a cloud-managed platform, host-level metrics can't be collected from the operating system the way they are for on-premises servers. Instead, Redgate Monitor retrieves them from the Google Cloud Monitoring API. To collect them you must provide Redgate Monitor with a Google Cloud service account that has the roles/monitoring.viewer role. To learn how to do that, see Adding Google Cloud SQL for SQL Server.
Unlike Amazon RDS, Google Cloud SQL has no separate "enhanced monitoring" option to enable — all of the machine metrics below are collected from the standard Google Cloud Monitoring API using your service account:
| Metric name | Google Cloud Monitoring metric |
|---|---|
Machine: Processor time | database/cpu/utilization |
Machine: Memory used | database/memory/utilization, database/memory/quota |
Network utilization | database/network/received_bytes_count, database/network/sent_bytes_count |
Disk used | database/disk/bytes_used |
Disk capacity | database/disk/quota |
Disk read and write bytes/sec | database/disk/read_bytes_count, database/disk/write_bytes_count |
Disk transfers/sec | database/disk/read_ops_count, database/disk/write_ops_count |
SQL Server instance metrics
These metrics are collected from the SQL Server instance itself using your SQL Server login, the same as for other monitored SQL Server instances.
- User connections
- SQL Server: total memory
- SQL Server: target memory
- SQL Server: free memory
- SQL Server: plan cache memory
- SQL Server: connection memory
- SQL Server: lock memory
- SQL Server: optimizer memory
- SQL Server: buffer cache memory
- SQL Server: granted workspace memory
- SQL Server: log pool memory
- SQL Server: reserved server memory
- SQL Server: maximum workspace memory
- SQL Server: stolen server memory
- Memory grants pending
- Batch requests/sec
- Compilations/sec
- Compilations/batch
- Recompilations/sec
- Buffer cache hit ratio
- Buffer page life expectancy
- Full scans/sec
- Page splits/sec
- Page splits/batch request
- Latch wait time
- Lock timeouts/sec
- Lock waits/sec
- Avg. lock wait time
SQL Server database metrics
- Transactions/sec
- Active transactions
- Total file size
- Total data file size
- Total data file used
- File size
- File used
- Total log file size
- Total log space used
- Log space used %
- Log bytes flushed/sec
- Log flushes/sec
- Log flush waits/sec
- Virtual log file count
Alerts
Alerts are based on metrics, so if the metric (or metrics) an alert uses are collected from the Google Cloud Monitoring API, that alert needs the Google Cloud service account in the same way the metric does. Machine-level alerts therefore require the service account, whereas SQL Server and database alerts are raised from data collected through your SQL Server login.
Both static (threshold-based) and dynamic (machine-learning) alerts are supported for Google Cloud SQL.
The following table describes the supported alerts:
| Alert name | Level raised | Google Cloud service account required? |
|---|---|---|
Backup overdue | SQL Server | - |
Blocking process | SQL Server | - |
Configuration change | SQL Server | - |
Custom metric | SQL Server | - |
Custom metric collection error | SQL Server | - |
Database file usage | SQL Server | - |
Database unavailable | SQL Server | - |
Differential backup overdue | SQL Server | - |
Disk space | Machine | Yes |
Fragmented indexes | SQL Server | - |
Internal Redgate Monitor error (instance) | SQL Server | - |
Internal Redgate Monitor error (host) | Machine | Yes |
Log backup overdue | SQL Server | - |
Long-running query | SQL Server | - |
Monitoring error (SQL Server data collection) | SQL Server | - |
Monitoring stopped (SQL Server credentials) | SQL Server | - |
Monitoring error (host data collection) | Machine | Yes |
Monitoring error (host credentials) | Machine | Yes |
Page verification | SQL Server | - |
Physical memory | Machine | Yes |
Processor under-utilization | Machine | Yes |
Processor utilization | Machine | Yes |
Server waits | SQL Server | - |
SQL Server error log entry | SQL Server | - |
SQL Server log entry | SQL Server | - |
SQL Server instance unreachable | SQL Server | - |
Suspended wait time | SQL Server | - |
Virtual log file count | Database | - |
Enterprise Tier alerts
The Enterprise Tier alerts available on some other platforms — Server role membership changes and SQL Audit file rollover missed event — are not available on Google Cloud SQL, because both rely on SQL Server Audit, which Redgate Monitor does not collect from Google Cloud SQL instances (see below).
What's not available on Google Cloud SQL
Because Google Cloud SQL is a managed platform that restricts access to some SQL Server features, the following are not available when monitoring a Google Cloud SQL for SQL Server instance:
- Query execution sampling — not available. Google Cloud SQL does not provide the Extended Events access Redgate Monitor uses to sample query executions.
- SQL Deadlock (XE) alert — not available, as it relies on Extended Events.
- SQL Server Agent alerts — Job cancelled, Job duration unusual, Job failing and SQL Server Agent Service status are not available, because Google Cloud SQL does not provide SQL Server Agent.
- SQL Audit-based alerts — SQL Audit file rollover missed event and Server role membership changes (both Enterprise Tier) are not available.
- Version store usage alert — not available for this instance type.
This documentation contains proprietary information and is protected by copyright law.
Copyright © 2026 Red Gate Software Limited. All rights reserved
