Executed Job stdout/stderr doesn't appear in the event ID modal....

Hi guys…

I have some jobs that in certain cases produce a lot of stdout lines. I’ve noticed that the STDOUT attribute for that event ID doesn’t show up in the modal, when we click in the stream textarea for that specific ID.

Is there a limit regarding stdout/stderr??

This is pretty important for us if we want to troubleshoot a deployment.

There’s any control of that limit?

Thanks,

-Cesar

Hey Cesar,
Currently, the UI shows up to 4000 events within the standard out pane on the UI. The UI gets this value as a header from the API. There is a setting that can change this on the python side of things:


# Returned in the header on event api lists as a recommendation to the UI
# on how many events to display before truncating/hiding
RECOMMENDED_MAX_EVENTS_DISPLAY_HEADER = 4000

but I’d highly recommend against bumping it up for the time being. It will most likely cause some intense memory usage when looking at long jobs to the point where a user’s AWX tab will crash. Longer term, the team is looking at tweaking the performance/behavior of this page for dealing with large jobs, and I apologize for the inconvenience this causes in the meantime!

As far as troubleshooting a particular job run, you can filter the standard out with the search bar above the pane, such that only events of a particular status, verbosity or host name are displayed. The key next to the search bar to find the particular parameters you can query on. You can refer to the [Tower documentation on the feature](http://Tower documentation on the feature) to learn more about how to utilize our search filtering syntax.

Said docs which I meant to link to: http://docs.ansible.com/ansible-tower/latest/html/userguide/search_sort.html

And forgot to answer:

I’ve noticed that the STDOUT attribute for that event ID doesn’t show up in the modal, when we click in the stream textarea for that specific ID.

The modal you get when clicking a line of standard out shows standard out of the ansible module, not standard out of the playbook run (i.e. what’s in the main pane).

FWIW, on the api side we will truncate the event data structure if it contains more than 700K of output. So if you have really large output it will get zeroed out before we store it to protect the system from automatically ingesting an extreme amount of data. It’s tunable so if you need to relax that a bit you can: https://github.com/ansible/awx/blob/devel/awx/settings/defaults.py#L171-L173