Vertex AI Vector Search: validate embedding dimensions and distance measure at deploy time

TL;DR

Also verify the Vertex AI API is enabled in the project, and monitor index creation and deployment logs: a failed deployment is usually the API, the network and firewall settings, or the dimension config, in that order.

Steps

1. [googlecloudplatform/cloud-networking-solutions]: validate embedding dimensions and distance measure types when creating the index. Distance measure types are DOTPRODUCTDISTANCE, COSINEDISTANCE, and L2SQUARED_DISTANCE; shard sizes come in SMALL, MEDIUM, and LARGE. Recommended models: gemini-embedding-001 at 768 dimensions. Also verify the Vertex AI API is enabled in the project, and monitor index creation and deployment logs: a failed deployment is usually the API, the network and firewall settings, or the dimension config, in that order.

When to use

You are seeing this: [googlecloudplatform/cloud-networking-solutions]: validate embedding dimensions and distance measure types when creating the index. Use this skill when you run into "Vertex AI Vector Search: validate embedding dimensions and distance measure at deploy time".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.

Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.

Edge cases

• Also verify the Vertex AI API is enabled in the project, and monitor index creation and deployment logs: a failed deployment is usually the API, the network and firewall settings, or the dimension config, in that order.

Source: Dimension and distance-measure mismatches between the index config and the embedding model surface as failed deployments or wrong results.

Verified resolution: [googlecloudplatform/cloud-networking-solutions]: validate embedding dimensions and distance measure types when creating the index. Distance measure types are DOTPRODUCTDISTANCE, COSINEDISTANCE, and L2SQUARED_DISTANCE; shard sizes come in SMALL, MEDIUM, and LARGE. Recommended models: gemini-embedding-001 at 768 dimensions. Also verify the Vertex AI API is enabled in the project, and monitor index creation and deployment logs: a failed deployment is usually the API, the network and firewall settings, or the dimension config, in that order.

Source URL: https://github.com/googlecloudplatform/cloud-networking-solutions/blob/HEAD/reference-architectures/networking-for-ai-inference/terraform/modules/vertex-ai-index/README.md