Add error to point user to slurm resume log

Author: hgreebe

src/slurm_plugin/instance_manager.py

@@ -848,7 +848,7 @@ def _best_effort_node_assignment(
             # No instances launched at all, e.g. CreateFleet API returns no EC2 instances,
             # or no left instances available from a best-effort EC2 launch
             logger.info("No launched instances found for nodes %s", print_with_count(nodes_resume_list))
-            self._update_failed_nodes(set(nodes_resume_list), "InsufficientInstanceCapacity", override=False)
+            self._update_failed_nodes(set(nodes_resume_list), "InsufficientInstanceCapacity(Check slurm_resume log for ec2 error codes)", override=False)
 
     def _all_or_nothing_node_assignment(
         self,
@@ -903,7 +903,7 @@ def _all_or_nothing_node_assignment(
             # No instances launched at all, e.g. CreateFleet API returns no EC2 instances,
             # or no left instances available from a best-effort EC2 launch
             logger.info("No launched instances found for nodes %s", print_with_count(nodes_resume_list))
-            self._update_failed_nodes(set(nodes_resume_list), "InsufficientInstanceCapacity", override=False)
+            self._update_failed_nodes(set(nodes_resume_list), "InsufficientInstanceCapacity(Check slurm_resume log for ec2 error codes)", override=False)
 
     def _launch_instances(  # noqa: C901
         self,
@@ -986,7 +986,7 @@ def _launch_ec2_instances(self, batch_nodes, compute_resource, fleet_manager, in
             #   queue_2: {cr_3: list[EC2Instance]}
             # }
         else:
-            self._update_failed_nodes(set(batch_nodes), "InsufficientInstanceCapacity")
+            self._update_failed_nodes(set(batch_nodes), "InsufficientInstanceCapacity(Check slurm_resume log for ec2 error codes)")
         return launched_ec2_instances
 
     def _get_fleet_manager(self, all_or_nothing_batch, compute_resource, queue):