System Insights — A GCP CloudSQL system monitoring tool

Published in

Fournine Cloud

4 min readOct 19, 2023

If you have CloudSQL instances running (MySQL or PostgreSQL) you might be aware of Query insights, which help developers to identify inefficiencies and address them. However, there is no single-page view for system-level metrics. In this blog, let’s discuss why we need system insights and what actually it offers.

Why System Insights?

While there is brief information on the instance overview page, it does not allow users to view and compare different metrics at the same time.

Personally, I know many organizations rely on third-party monitoring tools just to have a universal view of the metrics to get a better understanding of their CloudSQL instance. To make this easier, system insights are introduced.

System Insights

System Insights is a monitoring tool for database systems, that consolidates essential metrics, events, and logs to offer a holistic perspective on both external database performance and internal system resources. The integration of all these signals into a unified dashboard enables swift identification of potential problem sources, eliminating the need to toggle between different tools. Currently, System insights are accessible for Cloud SQL for PostgreSQL and Spanner in GA, and for Cloud SQL for MySQL in preview.

Key features of System Insights:

A quick snapshot of the system status
Pre-built dashboards
Correlation
Customizable views for special cases

A quick snapshot of the system status

On the System Insights page’s upper section, you’ll encounter a summary of your system’s primary assets and its present condition. This provides a rapid visual assessment of any possible concerns that might require your attention. The system setup enables you to gain insights into how well the system is operating in relation to the initially allocated resources.

Pre-built dashboards

Providing pre-built dashboards will help generalists and specialists and let them take the required action immediately. For example, system insights is providing a prebuilt dashboard called disk storage by type which will be handy to verify what is taking more storage and help to take needed actions.

Correlation (beyond metrics)

For effective troubleshooting of database issues, users often extend their analysis beyond mere metrics and delve into logs and system occurrences. For instance, a slow query lacking a proper index can lead to extended lock wait times, resulting in heightened error rates or increased CPU utilization, even when the external workload remains unchanged.

Comprehending the underlying causes behind such behaviors solely through metrics can be quite challenging. A proficient database expert will employ a cloud-based logging tool alongside system monitoring to identify any system events that coincide with an anomaly in resources. This process not only entails toggling between various tools but also entails extracting granular data from an extensive log repository that isn’t initially user-friendly.

With System insights, GCP has streamlined this process for users. Instead of requiring users to navigate between numerous tools, GCP seamlessly incorporated events into the System insights dashboard. Critical system operations are extracted from cloud logs and presented to users in a visual chronological format, known as the ‘events timeline.’ It offers a dynamic experience, enabling users to naturally traverse through time and establish correlations between metrics and system events.

Customizable views for special cases

The preset settings on the dashboard serve as an excellent initial configuration for the majority of users. However, for individuals seeking a higher degree of influence over the metrics they want to track, system insights offer customizable perspectives. Users have the flexibility to augment the preselected metrics or opt for a specific subset that aligns more closely with their requirements.

As System Insights widens its scope of telemetry coverage and introduces additional dimensions for customization, this functionality will cater to the evolving demands of database experts without overwhelming those who prefer the default experience designed for general users.

Let me know your favorite thing about system insights and I would like to know if you think any new features are needed to monitor CloudSQL.