Struct S3RemoteCacheConfig 

pub struct S3RemoteCacheConfig {
    pub bucket: String,
    pub prefix: String,
    pub namespace: String,
    pub region: String,
    pub endpoint: Option<Url>,
    pub force_path_style: Option<bool>,
    pub conditional_writes: S3ConditionalWrites,
    pub credentials: S3Credentials,
    pub connect_timeout: Duration,
    pub read_timeout: Duration,
    pub download_timeout: Duration,
    pub retries: i64,
}

Connection, addressing, and credential settings for an S3 remote cache.

Fields

bucket: String

Bucket holding the cache.

prefix: String

Key prefix within the bucket, empty for the bucket root.

namespace: String

Namespace isolating one project’s cache, used as a key prefix.

region: String

Region used to sign requests.

endpoint: Option<Url>

Endpoint override for a non-AWS store such as MinIO or R2.

force_path_style: Option<bool>

Force bucket-in-path addressing, overriding what the endpoint implies.

conditional_writes: S3ConditionalWrites

How to treat a store that does not implement conditional writes.

credentials: S3Credentials

Credentials used to sign requests.

connect_timeout: Duration

Maximum time allowed to establish a connection.

read_timeout: Duration

Maximum time without response progress for ordinary requests.

download_timeout: Duration

Deadline for one blob download, spanning every retry attempt and the backoff between them rather than bounding a single attempt.

A single stalled attempt is already bounded by connect_timeout and read_timeout, so this budget exists to cap the total wall-clock one logical download may spend: exhausting it fails the download even when retries remain. Size it for the largest artifact worth waiting on, not for one attempt at it.

retries: i64

Number of attempts after the initial request for retryable failures.