A $9,400 GCP bill where $3,000 had no owner

The GCP account requiring the most extensive optimization for the reporter was not the highest-cost one, but rather the most fragmented among eleven projects. These projects lacked consistent naming, resource labels, and a billing export setup. A total of $9,400 was spent monthly across the organization, but without visibility into where the money was going, $3,000 remained unaccounted for initially.

Upon starting the optimization process, the reporter spent a week implementing instrumentation, turning on billing export to BigQuery, applying resource labels, documenting project ownership, and setting budget alerts at both the organization and project levels. This single week of work transformed the conversation around spend, allowing for more specific questions.

For instance, Project 7 was found to be using $1,400 monthly, which belonged to a contractor who had already left the company. Three Compute Engine instances and a Cloud SQL database were identified as the culprits, but their purpose remained unknown. The reporter paused these resources and observed that nothing critical was affected. They were subsequently deleted, resulting in an immediate saving of $1,400.

The reporter emphasizes that the fragmentation itself constituted a cost, encompassing engineer hours spent grappling with the account's confusion, slow incident response due to uncertain project identification, and the inability to make informed architectural decisions. They stress that visibility in a multi-project GCP environment is not a luxury, but rather the foundational element upon which all other optimizations are built.