[pd-router] Add Configurable Retry Logic for reduce backend pressure (#8744)

This commit is contained in:
Simo Lin
2025-08-04 20:42:07 -07:00
committed by GitHub
parent d98a4913ea
commit 354ac43555
10 changed files with 502 additions and 293 deletions

View File

@@ -4,7 +4,7 @@ use common::mock_worker::{HealthStatus, MockWorker, MockWorkerConfig, WorkerType
use futures_util::StreamExt;
use reqwest::Client;
use serde_json::json;
use sglang_router_rs::config::{PolicyConfig, RouterConfig, RoutingMode};
use sglang_router_rs::config::{PolicyConfig, RetryConfig, RouterConfig, RoutingMode};
use sglang_router_rs::routers::{RouterFactory, RouterTrait};
use std::sync::Arc;
@@ -36,6 +36,7 @@ impl TestContext {
request_id_headers: None,
max_concurrent_requests: 64,
cors_allowed_origins: vec![],
retry: RetryConfig::default(),
};
let mut workers = Vec::new();