NominalLogStreamOpts#

#[non_exhaustive] pub struct NominalLogStreamOpts#

Limits include ready and in-flight batches, so a stalled backend applies backpressure. Uploads use the same Conjure HTTP client and retry policy as time-series streaming. Construct with Default and customize its public fields before passing to the builder.

Fields#

pub max_request_bytes: usize#

Maximum uncompressed protobuf request size, including all framing.

pub max_buffered_bytes: usize#

Encoded size of all records held in memory: pending, queued or in flight. Each record counts its own bytes; shared channel and request framing is not counted. Enqueue blocks while this is full.

pub max_records_per_batch: usize#

Maximum number of records in one request.

pub max_request_delay: Duration#

Maximum time a record waits in a partial batch before upload.

pub num_upload_workers: usize#

Number of threads that upload batches in parallel.

pub base_api_url: String#
pub user_agent: ProductAgent#

Program name and version at the start of the User-Agent header of every request the log stream sends. See ProductAgent.

Trait implementations#

impl Default for NominalLogStreamOpts
fn default() -> Self

Derived: Clone, Debug

Auto traits: Send, Sync, Unpin, UnwindSafe, RefUnwindSafe