osproxy-sink 1.0.1

Write sink: Sink trait + OpenSearchSink now; QueueSink (Kafka) redundancy later behind the same trait.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

//! Write sink.
//!
//! Where writes go, isolated from how routing is decided
//! (`docs/decisions/008`). The [`Sink`] trait is the seam: `OpenSearchSink`
//! writes directly to a cluster, and the future queue-based redundancy mode is a
//! `QueueSink` drop-in behind the same trait. Epoch stamping is carried on every
//! [`WriteOp`] at this boundary (`docs/06` ยง2).
//!
//! M1 ships the [`Sink`] trait, the [`WriteBatch`]/[`WriteAck`] vocabulary, an
//! in-memory [`MemorySink`] for tests and dry-run, and the [`OpenSearchSink`]
//! that writes directly to a cluster over a pooled HTTP connection. M2 adds the
//! [`Reader`] seam for get-by-id reads (kept separate because a read is always
//! direct-to-cluster, a redundancy `QueueSink` can absorb writes but cannot
//! answer a read); both `MemorySink` and `OpenSearchSink` implement it.
#![deny(missing_docs)]

mod ack;
mod batch;
mod breaker;
mod conn;
mod error;
mod memory;
mod opensearch;
mod read;
mod sink;
mod trace_headers;
mod wire;

pub use ack::{OpResult, WriteAck};
pub use batch::{DocOp, WriteBatch, WriteOp};
pub use conn::PoolStats;
pub use error::SinkError;
pub use memory::MemorySink;
pub use opensearch::{buffered, stream_body, BodyError, ByteBody, OpenSearchSink};
pub use read::{
    CountOutcome, CursorOp, CursorOutcome, ForwardOp, ReadOp, ReadOutcome, Reader, SearchOp,
    SearchOutcome, StreamingForward, StreamingSearch,
};
pub use sink::Sink;