Fix race create sessione su CPX: ready solo dopo MediaMTX.

I nodi cloud restano in provisioning finché :9997 risponde; retry su create_path e 503 retryable se l’ingest è ancora irraggiungibile.

Co-authored-by: Cursor <cursoragent@cursor.com>

backend/app/services/streams/autoscaler.rb

@@ -136,6 +136,8 @@ module Streams
 
     def reconcile!
       actions = []
+      actions.concat(promote_provisioning_nodes!)
+      actions.concat(reclaim_stuck_provisioning!)
       m = self.class.metrics
 
       if need_capacity?(m) && can_provision?(m)
@@ -170,6 +172,30 @@ module Streams
 
     private
 
+    def promote_provisioning_nodes!
+      actions = []
+      StreamNode.where(status: "provisioning").find_each do |node|
+        next unless Streams::NodeHealth.promote_if_healthy!(node)
+
+        actions << :"ready_#{node.slug}"
+        Rails.logger.info("[Streams::Autoscaler] promoted #{node.slug} to ready")
+      end
+      actions
+    end
+
+    def reclaim_stuck_provisioning!
+      actions = []
+      stuck_after = ENV.fetch("STREAM_NODE_PROVISIONING_STUCK_MINUTES", "15").to_i.minutes.ago
+      StreamNode.where(status: "provisioning").where("created_at < ?", stuck_after).find_each do |node|
+        @provisioner.decommission!(node)
+        actions << :"reclaim_#{node.slug}"
+        Rails.logger.warn("[Streams::Autoscaler] decommissioned stuck provisioning #{node.slug}")
+      rescue NodeProvisioner::BusyError, NodeProvisioner::Error => e
+        Rails.logger.warn("[Streams::Autoscaler] reclaim #{node.slug}: #{e.message}")
+      end
+      actions
+    end
+
     def need_capacity?(m)
       m[:free_slots] <= self.class.soft_free_slots
     end