We are investigating a significant and unexpected spike in API usage associated with our Google Cloud project:
The project has been suspended due to suspected abusive activity consistent with potential API key compromise or hijacking. During the impacted period (March 21–23, 2026), we observed a substantial increase in usage and charges, primarily driven by:
-
Gemini API (image generation – high token output)
-
Places API (new usage observed)
At this time, we are attempting to perform a deep-dive investigation to determine:
-
The origin of the API calls (source IPs, regions)
-
Request metadata (headers, user agents, authentication method)
-
Whether the activity was authenticated vs unauthenticated
-
Any indicators of API key misuse or external abuse