FFAStrans forum

@ArsenioV
Hmmm sorry i might have misread your problem description.
You say you reinstalled the service but you don't mention if you set the needed credentials of the service that allows the service to browse UNC paths. (i guess you dont have "everyone read" permissions on your UNC locations?)

Easy test: don't start the webinterface as service but as local application. Is it able to browse your configured UNC paths? If yes, you just need to equip the windows service with some username that can list files on the UNC paths.
One more level deep: log on to windows using the exact same username that you want your service to run as in the final configuration and try to open a windows explorer to the configured UNC path. It must open without asking for password.

deleted

@ArsenioV
so it also does not work when you run server.exe as application (with your current credentials) instead of service?
I really need to see whats going on, could you do me the favour and run as application, attempt to open the file browser webpage in your browser, wait for the error to appear and PM me the last 50-100 lines from the console where server.exe is running?

@Silicon
Fixed Version is out, let me know if you face anymore troubles. (don't worry about the size of the exe, it will get bigger again later )

@ArsenioV
I tested the browsing on UNC again, didnt find any issues... need the requested log!

@emcodem
I've installed version 1.2.0.1 and I can confirm that both reported issues are fixed.
Thanks a lot!

Hi emcodem
Unfortunatelly I have found another bug in Webinterface 1.2.0.1: after a while (I don't know how long it takes) the list of Finished jobs in bottom part of page becames unsorted ... it should be sorted based on End time (correct?), but it is not.

Uhhh... lemme see if i can reproduce that.
The worst issues are the ones that take "time" to be visible ^^

Hi,
I've been watching the Webinterface page live now and it took aprox. 20 min. (and aprox. 30+ jobs) to get the list of Finished jobs wrongly sorted. Number of jobs processed seems to be more important that time elapsed.
I have forgotten to mention, that F5 fixes it ... but this is obvious

Hey @Silicon

i am really greatful that you submit all the anomalies you find to us, thanks a lot for it!
It was really easy to reproduce by just submitting lots of jobs for about 10 minutes, pretty obvious. I fear it will still take some time because each "test" costs again about 10 minutes but i will for sure solve this, no worries! (it's totally disturbing from user perspective^^)

@Silicon @ernie1901
Fixed, sorry for the delay. Please check out release 1.2.0.2

On my way on adding sophisticated monitoring, some metrics can now be polled for Prometheus/Influxdb or similar systems.
Users that have such a system already in place might want to try it out. All others need to wait until i finished the prometheus integration in webinterface

http://localhost:3003/metrics

Always reported:
ffas_jobs_incoming_count COUNT
ffas_jobs_queue_count COUNT
ffas_jobs_running_count COUNT

while a job is running, it's runtime in seconds along with wf_name and hostname is reported
ffas_running_job {job_id=JOBID,wf_name=WFNAME,hostname=HOST} seconds

As soon as a job ended with error, it should be reported ONCE too but this is currently buggy as ALL job finished are reported as error
ffas_jobs_error_job_info {job_id=JOBID,wf_name=WFNAME,file=FILENAME,status=MESSAGE} 1

Using those metrics, when configured correctly, the job/branch runtime can be displayed on grafana or similar systems like this: