This is lowish priority since users can always pull their logs directly from S3. The service endpoint just eases getting the logs, especially for CTS client users.
Currently logs are only available after a job goes through the error processing submitting -> error processing -> error transitions. However, logs for multi-container jobs on the kbase site are available as soon as the container reports the error state to the service, which may be much sooner than when the parent job transitions to error.
- Make the logpath always available.
- Add a container state user endpoint that returns whether logs are available. Also could subsume the exit codes method into this endpoint.
This is lowish priority since users can always pull their logs directly from S3. The service endpoint just eases getting the logs, especially for CTS client users.
Currently logs are only available after a job goes through the error processing submitting -> error processing -> error transitions. However, logs for multi-container jobs on the
kbasesite are available as soon as the container reports the error state to the service, which may be much sooner than when the parent job transitions to error.