fix(clusters): Add last_scheduling_status to errors

Signed-off-by: Kyle Squizzato <kyle@replicated.com>

Author: squizzi

package-lock.json

@@ -4,6 +4,7 @@ export class Cluster {
   name: string;
   id: string;
   status: string;
+  last_scheduling_status?: string;
 }
 
 export class ClusterVersion {
@@ -177,9 +178,9 @@ export async function pollForStatus(vendorPortalApi: VendorPortalApi, clusterId:
         return clusterDetails;
       }
 
-      // Once state is "error", it will never change. So we can shortcut polling.
       if (clusterDetails.status === "error") {
-        throw new Error(`Cluster has entered error state.`);
+        const schedulingStatus = clusterDetails.last_scheduling_status ? `, last scheduling status: ${clusterDetails.last_scheduling_status}` : "";
+        throw new Error(`Cluster has entered error state${schedulingStatus}`);
       }
 
       console.debug(`Cluster status is ${clusterDetails.status}, sleeping for ${sleeptimeMs / 1000} seconds`);
@@ -219,7 +220,8 @@ async function getClusterDetails(vendorPortalApi: VendorPortalApi, clusterId: st
   return {
     name: body.cluster.name,
     id: body.cluster.id,
-    status: body.cluster.status
+    status: body.cluster.status,
+    last_scheduling_status: body.cluster.last_scheduling_status
   };
 }