[Test] Add automated root cause analysis for cluster creation failure.

[gmarciani]

Description of changes

Add automated root cause analysis for cluster creation failure.
If cluster creation fails, the test now prints out the last 10 errors from the most critical logs and attempt to identify a root cause from them.
For the time being, the only root cause that we extract is the ICE.

Here is the sample diagnosis report printed by a test failure:

=== ROOT CAUSE ANALYSIS ===

--- SUMMARY ---
['InsufficientInstanceCapacity on p4d.24xlarge']

--- /var/log/chef-client.log ---
[2026-02-26T19:59:41+00:00] ERROR: shard_seed: Failed to get dmi property serial_number: is dmidecode installed?
... other 10 errors ...

--- /var/log/parallelcluster/clustermgtd ---
2026-02-26 20:25:14,163 - [slurm_plugin.fleet_manager:_launch_instances] - ERROR - Failed RunInstances request (3fd67297-4cbc-4d20-94b3-370e060495a6): InsufficientInstanceCapacity - We currently do not have sufficient p4d.24xlarge capacity in the Availability Zone you requested (us-east-1a). Our system will be working on provisioning additional capacity. You can currently get p4d.24xlarge capacity by not specifying an Availability Zone in your request or choosing us-east-1b, us-east-1c, us-east-1d.
... other 10 errors ...

--- /var/log/cfn-init.log ---
DEBUG:root:CustomActionsConfig(stack_name='integ-tests-u8lflx6fp7yct6bp-integ-mg', cluster_name='integ-tests-u8lflx6fp7yct6bp-integ-mg', region_name='us-east-1', node_type='HeadNode', queue_name='', pool_name='', resource_name='', instance_id='i-0c94f0e998cdd9f63', instance_type='c5.18xlarge', ip_address='192.168.3.21', hostname='ip-192-168-3-21.ec2.internal', availability_zone='us-east-1a', scheduler='slurm', event_name='OnNodeStart', legacy_event=<LegacyEventName.ON_NODE_START: 'preinstall'>, can_execute=False, dry_run=False, script_sequence=[], script_sequences_per_event={<LegacyEventName.ON_NODE_START: 'preinstall'>: [], <LegacyEventName.ON_NODE_CONFIGURED: 'postinstall'>: [], 
... other 10 errors ...

--- /var/log/cloud-init-output.log ---
2026-02-26 19:59:05,767 - util.py[WARNING]: Running module selinux (<module 'cloudinit.config.cc_selinux' from '/usr/lib/python3.9/site-packages/cloudinit/config/cc_selinux.py'>) failed
... other 10 errors ...

--- /var/log/parallelcluster/slurmctld.log ---
No error found

Tests

Executed integ test test_efa known that it is going to face ICE. The test report shows the expected diagnosis (see above)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.