Cloud Recommender is a managed service that analyzes resource usage across your environment and provides suggestions to help you optimize your infrastructure. It is built from a catalog of individual recommenders, where each recommender targets a specific product and resource type and looks at a particular dimension of how those resources are being used. The Professional Cloud Database Engineer exam treats Cloud Recommender as the standard way Google Cloud surfaces optimization suggestions, so the questions tend to focus on which recommender solves a given problem rather than on the mechanics of the service.
Two behaviors are worth being precise about. Cloud Recommender does not apply recommendations automatically, so you keep full control over your environment and decide whether to act on a suggestion. To make acting on one straightforward, each recommendation provides a link to apply it directly to the service. Those two points, no automatic changes and a direct link to apply, describe how the service is meant to be used.
The recommenders are grouped into categories that map to common goals: cost, security, performance, reliability, and manageability. There is no need to memorize the entire catalog. What helps for the Professional Cloud Database Engineer exam is recognizing which category a scenario falls into and which recommender addresses it, particularly for Cloud SQL, which has more recommenders than any other service in the catalog.
Under cost, the Cloud SQL overprovisioned instance recommender evaluates historical CPU and memory metrics and suggests smaller, more cost-effective machine types. This is the one most likely to appear on the exam as a solution for reducing database spend, so it is worth knowing by name. The Cloud SQL idle instance recommender belongs to the same category and identifies databases that are not being used, which is another way to cut cost.
In the security category, the Identity and Access Management recommender identifies excess permissions and suggests more restrictive roles, which supports least-privilege access. The Cloud SQL security recommender flags risks such as public IP addresses on an instance.
For performance, the Cloud SQL performance recommender analyzes database flags and settings and points out adjustments. The VM machine type recommender works in the other direction from the overprovisioned recommender and suggests upgrades for instances running at high utilization.
Under reliability, the Cloud SQL reliability recommender helps with high availability configurations, and the Cloud SQL out-of-disk recommender warns you before a database runs out of storage space. For manageability, the deprecation and breaking changes recommender alerts you to upcoming API retirements, and the Error Reporting notification recommender helps you stay on top of application crashes.
For the database-focused scenarios you are likely to see, the useful pattern is to read the goal in the question and match it to a category. A prompt about lowering the bill for an underused Cloud SQL instance points toward the cost recommenders, the overprovisioned instance recommender in particular. A prompt about an instance exposed on a public IP points toward the security recommender. A warning about running out of space points toward the out-of-disk recommender under reliability. Keeping the category-to-goal mapping straight, and remembering that the service suggests rather than applies, covers most of what the Professional Cloud Database Engineer exam asks about Cloud Recommender.
Our Professional Cloud Database Engineer course covers Cloud Recommender alongside Cloud SQL machine sizing and high availability, with practice questions that drill these distinctions.
