fix(kiro,pool,model): 对齐 Kiro 管理链路并修复全局模型删除行为 (#305)

* feat(pool): 号池支持跳过额度耗尽账号

- 新增 pool_advanced.skip_exhausted_accounts 开关及高级设置 UI, 默认关闭并兼容旧配置
- 为 Codex/Kiro 增加额度耗尽判定, 接入请求侧候选跳过并新增 account_quota_exhausted skip reason
- 号池列表将额度耗尽账号标记为 blocked/额度耗尽, 并补充前后端相关测试

* fix(kiro): 对齐账号管理与 provider-query 的 Rust 行为

- 修复 Kiro 单条导入误走 import-refresh-token 的前端分流, 并为误用路径返回明确错误提示
- 为 Kiro 导入与本地请求链补齐 bearer 兼容, 同步放开账号启停等 Key 更新操作的 auth_type 校验
- 实现 Kiro provider-query 本地模型测试与 failover 执行链, 并修复结果弹窗在无 trace 时无法展示 attempts/响应体的问题

* fix(model): 删除全局模型时级联清理关联提供商模型

- 对齐 Python 版本删除逻辑, GlobalModel 删除前先在事务内清理关联的 Provider Model 记录
- 修复已绑定 Provider 的模型在 Rust SQL 仓库下会被外键约束拦住、无法正常删除的问题
- 增加管理端回归测试, 覆盖绑定 Provider Model 的 GlobalModel 删除场景

* fix(kiro,ci): 恢复 Kiro OAuth 持久化并修复 Rust CI

* Fix oauth-managed provider key semantics

---------

Co-authored-by: fawney19 <[email protected]>

apps/aether-gateway/src/scheduler/candidate/runtime.rs

@@ -1,5 +1,6 @@
 use std::collections::{BTreeMap, BTreeSet};
 
+use aether_admin::provider::pool as admin_provider_pool_pure;
 use aether_data_contracts::repository::candidates::StoredRequestCandidate;
 use aether_data_contracts::repository::provider_catalog::StoredProviderCatalogKey;
 use aether_scheduler_core::{
@@ -20,6 +21,7 @@ pub(super) struct CandidateRuntimeSelectionSnapshot {
     pub(super) provider_concurrent_limits: BTreeMap<String, usize>,
     pub(super) provider_key_rpm_states: BTreeMap<String, StoredProviderCatalogKey>,
     provider_quota_blocks_requests: BTreeMap<String, bool>,
+    key_account_quota_exhausted: BTreeMap<String, bool>,
     provider_key_rpm_reset_ats: BTreeMap<String, Option<u64>>,
 }
 
@@ -30,7 +32,14 @@ pub(super) async fn read_candidate_runtime_selection_snapshot(
 ) -> Result<CandidateRuntimeSelectionSnapshot, GatewayError> {
     let recent_candidates = state.read_recent_request_candidates(128).await?;
     let provider_concurrent_limits = read_provider_concurrent_limits(state, candidates).await?;
+    let provider_skip_exhausted_accounts =
+        read_provider_skip_exhausted_account_map(state, candidates).await?;
     let provider_key_rpm_states = read_provider_key_rpm_states(state, candidates).await?;
+    let key_account_quota_exhausted = read_key_account_quota_exhaustion_map(
+        candidates,
+        &provider_key_rpm_states,
+        &provider_skip_exhausted_accounts,
+    );
     let provider_quota_blocks_requests =
         read_provider_quota_block_map(state, candidates, now_unix_secs).await?;
     let provider_key_rpm_reset_ats =
@@ -41,6 +50,7 @@ pub(super) async fn read_candidate_runtime_selection_snapshot(
         provider_concurrent_limits,
         provider_key_rpm_states,
         provider_quota_blocks_requests,
+        key_account_quota_exhausted,
         provider_key_rpm_reset_ats,
     })
 }
@@ -89,6 +99,11 @@ pub(super) fn is_candidate_selectable(
             .get(candidate.provider_id.as_str())
             .copied()
             .unwrap_or(false),
+        account_quota_exhausted: snapshot
+            .key_account_quota_exhausted
+            .get(candidate.key_id.as_str())
+            .copied()
+            .unwrap_or(false),
         rpm_reset_at: snapshot
             .provider_key_rpm_reset_ats
             .get(candidate.key_id.as_str())
@@ -122,6 +137,11 @@ pub(super) fn current_candidate_runtime_skip_reason(
         now_unix_secs,
         cached_affinity_target,
         provider_quota_blocks_requests,
+        account_quota_exhausted: snapshot
+            .key_account_quota_exhausted
+            .get(candidate.key_id.as_str())
+            .copied()
+            .unwrap_or(false),
         rpm_reset_at,
     })
 }
@@ -192,6 +212,64 @@ async fn read_provider_quota_block_map(
     Ok(quota_blocks)
 }
 
+async fn read_provider_skip_exhausted_account_map(
+    state: &(impl SchedulerRuntimeState + ?Sized),
+    candidates: &[SchedulerMinimalCandidateSelectionCandidate],
+) -> Result<BTreeMap<String, bool>, GatewayError> {
+    let provider_ids = candidates
+        .iter()
+        .map(|candidate| candidate.provider_id.clone())
+        .collect::<BTreeSet<_>>()
+        .into_iter()
+        .collect::<Vec<_>>();
+    if provider_ids.is_empty() {
+        return Ok(BTreeMap::new());
+    }
+
+    let providers = state
+        .read_provider_catalog_providers_by_ids(&provider_ids)
+        .await?;
+    Ok(providers
+        .into_iter()
+        .map(|provider| {
+            let skip_exhausted_accounts = provider
+                .config
+                .as_ref()
+                .and_then(|value| value.get("pool_advanced"))
+                .and_then(serde_json::Value::as_object)
+                .and_then(|value| value.get("skip_exhausted_accounts"))
+                .and_then(serde_json::Value::as_bool)
+                .unwrap_or(false);
+            (provider.id, skip_exhausted_accounts)
+        })
+        .collect())
+}
+
+fn read_key_account_quota_exhaustion_map(
+    candidates: &[SchedulerMinimalCandidateSelectionCandidate],
+    provider_key_rpm_states: &BTreeMap<String, StoredProviderCatalogKey>,
+    provider_skip_exhausted_accounts: &BTreeMap<String, bool>,
+) -> BTreeMap<String, bool> {
+    candidates
+        .iter()
+        .map(|candidate| {
+            let exhausted = provider_skip_exhausted_accounts
+                .get(candidate.provider_id.as_str())
+                .copied()
+                .unwrap_or(false)
+                && provider_key_rpm_states
+                    .get(candidate.key_id.as_str())
+                    .is_some_and(|key| {
+                        admin_provider_pool_pure::admin_pool_key_account_quota_exhausted(
+                            key,
+                            candidate.provider_type.as_str(),
+                        )
+                    });
+            (candidate.key_id.clone(), exhausted)
+        })
+        .collect()
+}
+
 fn read_provider_key_rpm_reset_at_map(
     state: &(impl SchedulerRuntimeState + ?Sized),
     candidates: &[SchedulerMinimalCandidateSelectionCandidate],