This document describes concepts related to the
gcp-starter environment. To learn about indexes and other environments, see Understanding.
Unlike other Starter Plan regions, the
gcp starter region has unique features and limitations.
Like other Starter Plan environments, projects on the
gcp-starter environment support one pod with enough resources to support approximately 100,000 vectors with 1536-dimensional embeddings and metadata; the capacity is proportional for other dimensions. Indexes on the
gcp-starter environment do not specify pod types;
create_index calls ignore the
Unlike other Starter Plan environments, projects in the
gcp-starter region have no retention limits; data is retained indefinitely. Indexes in these projects are not deleted after inactivity as in other environments.
After upgrading from the Starter Plan, you keep your free
gcp-starter project: you are not charged for use on this project. To use your data on a Standard or Enterprise environment, insert the data to a new project in a supported environment.
Like other Starter Plan environments, projects in the
gcp-starter environment do not support replicas.
After upserting records to indexes in
describe_index_stats operations may not return updated records for up to 10 seconds.
gcp-starter environment does not support the following features:
- Delete by metadata
configure_index(). Because this environment only supports one index configuration, this operation is not supported.
describe_index()with metadata filtering
deleteAllparameter. This feature is dependent on namespaces.
- Metadata strings containing the character
Δ. This is a reserved character in the underlying storage format.
Because projects on the
gcp-starter region do not support the above features, you may need to use different features when developing your project.
Because projects in the
gcp-starter environment do not support namespaces, you may wish to use metadata filtering to support multitenancy in your project. Another alternative is to upgrade to the Standard or Enterprise plans and use multiple indexes to support multitenancy.
Projects in the
gcp-starter environment do not support the collections feature. However, collections may not be necessary or appropriate for projects in the
gcp-starter environment. Collections serve two primary purposes: decreasing usage by archiving inactive indexes, and experimenting with different index configurations. However, projects on the
gcp-starter environment neither incur usage costs nor specify pod types or sizes.
Projects in the
gcp-starter environment do not support deleting records by metadata. In some cases, you may be able to delete records by ID instead. This may require you to first query your index with metadata filters, then extract the record IDs.
Updated about 2 months ago