feat: proactive rate limiting (RPM) + 429 short cooldown for sources

- config.go: Source add RPM field (requests-per-minute cap, 0=unlimited)
- provider.go: RecordRateLimit() — 429 uses fixed 30s cooldown, not exponential
- provider.go: Throttle() — token-bucket proactive rate limiter, spaces requests
  at 60s/RPM interval, respects context cancellation
- provider.go: ReportStatus() — 429 -> RecordRateLimit, 5xx -> RecordFailure
- provider.go: Chat/ChatStream — wire Throttle after TryAcquire
- api.go: sourcePayload + RPM, buildSource passes RPM through
- ui/index.html: add RPM input field in source editor, bilingual i18n labels
- deploy.sh: backup old binary + rollback on healthcheck failure
- provider_test.go: TestModelStateRateLimitShortCooldown, TestThrottleSpacingAndCancel
- config.yaml: sensenova rpm: 12

internal/gateway/api.go

@@ -72,6 +72,7 @@ type sourcePayload struct {
 	Temperature   float64                `json:"temperature"`
 	MaxTokens     int                    `json:"max_tokens"`
 	MaxConcurrent int                    `json:"max_concurrent"`
+	RPM           int                    `json:"rpm"` // optional requests-per-minute cap, 0 = unlimited
 }
 
 func (g *Gateway) handleSourcesAPI(w http.ResponseWriter, r *http.Request) {
@@ -105,6 +106,7 @@ func (g *Gateway) handleSourcesAPI(w http.ResponseWriter, r *http.Request) {
 			Temperature:   p.Temperature,
 			MaxTokens:     p.MaxTokens,
 			MaxConcurrent: p.MaxConcurrent,
+			RPM:           p.RPM,
 		}
 		if err := g.core.AddSource(src); err != nil {
 			writeError(w, http.StatusBadRequest, "source_error", err.Error())

internal/gateway/ui/index.html

@@ -847,6 +847,7 @@
           mEp: "聊天端点",
           mImgEp: "生图端点",
           mConc: "并发上限",
+          mRPM: "RPM 限速 (0=不限)",
           mTemp: "温度",
           mModels: "模型列表",
           mAddModel: "+ 模型",
@@ -1067,6 +1068,7 @@
           mEp: "Chat endpoint",
           mImgEp: "Image endpoint",
           mConc: "Max concurrency",
+          mRPM: "RPM limit (0 = unlimited)",
           mTemp: "Temperature",
           mModels: "Models",
           mAddModel: "+ model",
@@ -2479,6 +2481,10 @@
         <div><label>${t("mConc")}</label><input id="s-conc" type="number" value="${s.max_concurrent || 8}"></div>
         <div><label>${t("mTemp")}</label><input id="s-temp" type="number" step="0.1" value="${s.temperature ?? 0.7}"></div>
       </div>
+      <div class="row">
+        <div><label>${t("mRPM")}</label><input id="s-rpm" type="number" min="0" placeholder="0 = 不限" value="${s.rpm || 0}"></div>
+        <div></div>
+      </div>
       <label>${t("mModels")}</label>
       <div id="s-models"></div>
       <button class="ghost small" onclick="addModelRow()">${t("mAddModel")}</button>
@@ -2537,6 +2543,7 @@
           endpoint: $("#s-ep").value.trim(),
           image_endpoint: $("#s-img").value.trim(),
           max_concurrent: parseInt($("#s-conc").value) || 8,
+          rpm: parseInt($("#s-rpm").value) || 0,
           temperature: parseFloat($("#s-temp").value) || 0,
           models,
           meta,