Hi

we run a Galaxy instance (16.04) in multi-thread mode ( 8 x web handler / 2 jobs handler) to manage scalability, with Apache in frontend, and a cluster (SGE like).

We test the load of our infrastructure by running several workflows in simultaneous : 50 workflow with 35 steps (light cpu consuming for each steps)

All jobs are successfull but some steps remain in grey (it seems jobs are really ended because the next streps runs), before to appear in green (no yellow).

If we force the refresh history (refresh button on the interface), jobs appear in green.

When we run only one (or 2:)) workflow, each workflow follow the chain : grey, yellow, green.

Maybe, there is some mechanisms to optimize between web handler/apache and clients to avoid this and to get the right color status ?

Any clue ?

Thanks