Integration with Amazon CloudWatch Logs
Starting with AWS ParallelCluster version 2.6.0, common logs are stored in CloudWatch Logs by default. For more information about CloudWatch Logs, see Amazon CloudWatch Logs User Guide. To configure CloudWatch Logs integration, see the [cw_log] section and the cw_log_settings setting.
A log group is created for each cluster with a name
/aws/parallelcluster/
(for example,
cluster-name
/aws/parallelcluster/testCluster
). Each log (or set of logs if the path contains a *
) on
each node has a log stream named
.
(For example {hostname}
.{instance_id}
.{logIdentifier}
ip-172-31-10-46.i-02587cf29cc3048f3.nodewatcher
.) Log data is sent to CloudWatch by the CloudWatch agent, which
runs as root
on all cluster instances.
Starting with AWS ParallelCluster version 2.10.0, an Amazon CloudWatch dashboard is created when the cluster is created. This dashboard makes it easy to review the logs stored in CloudWatch Logs. For more information, see Amazon CloudWatch dashboard.
This list contains the logIdentifier
and path for the log streams available for platforms, schedulers, and nodes.
Platforms | Schedulers | Nodes | Log streams |
---|---|---|---|
amazon centos ubuntu |
awsbatch slurm |
HeadNode |
dcv-authenticator: dcv-ext-authenticator: dcv-agent: dcv-xsession: dcv-server: dcv-session-launcher: Xdcv: cfn-init: chef-client: |
amazon centos ubuntu |
awsbatch slurm |
ComputeFleet HeadNode |
cloud-init: supervisord: |
amazon centos ubuntu |
slurm |
ComputeFleet |
cloud-init-output: computemgtd: slurmd: |
amazon centos ubuntu |
slurm |
HeadNode |
clustermgtd: slurm_resume: slurm_suspend: slurmctld: |
amazon centos |
awsbatch slurm |
ComputeFleet HeadNode |
system-messages: |
ubuntu |
awsbatch slurm |
ComputeFleet HeadNode |
syslog: |
Jobs in clusters that use AWS Batch store the output of jobs that reached a RUNNING
,
SUCCEEDED
, or FAILED
state in CloudWatch Logs. The log group is /aws/batch/job
, and the
log stream name format is
. By
default, these logs are set to never expire, but you can modify the retention period. For more information, see Change log data retention in CloudWatch Logs in the
Amazon CloudWatch Logs User Guide.jobDefinitionName
/default/ecs_task_id
Note
chef-client
, cloud-init-output
, clustermgtd
, computemgtd
,
slurm_resume
, and slurm_suspend
were added in AWS ParallelCluster version 2.9.0. For
AWS ParallelCluster version 2.6.0, /var/log/cfn-init-cmd.log
(cfn-init-cmd
) and
/var/log/cfn-wire.log
(cfn-wire
) were also stored in CloudWatch Logs.