RSP portal background monitor and email


When submitting queries through the portal I occasionally send it to background.

It seems that sometimes the monitor never shows the job as completing but when I click on the info icon the job has actually finished

Also I sometimes provide an email address but I’ve never received an email.
Am I doing something wrong or mis-understanding the process?

Thanks for your question, @MRead!

I have confirmed that I, too, do not get an e-mail alert when my Portal background job is completed, even though I have enabled e-mail notification and included my e-mail address. This may be a bug in the current interface.

So far, I have not had any issues with the Portal’s background monitor not showing the job as completing, but, admittedly, I have not exercised this feature extensively myself. Does this problem seem to occur more often for you for certain types of queries (e.g… particularly long-running queries)?


1 Like

Email notifications are not currently part of the RSP architecture. For the RSP as a whole, not just the Portal, we recognize that some sort of notification system is needed, especially given the expectation that some Qserv full-Object-catalog queries will run for long times, and it is on the RSP development roadmap to select a notification approach and implement it. We will consider alternatives to email at that time that may integrate better into contemporary workflows.

In the mean time, it was simply an oversight that that “enable email” UI element was left visible in the Portal, and we’ll fix that.

The other observation about the background monitor occasionally missing a state change in the underlying query has been made within the team - I do run into it myself - and by users before. There are some complex network-connection-lifetime issues that couple into this. We have a revision of the background monitor to generalize it from just catalog queries to all long-running backgrounded jobs (e.g., from UWS services) that is planned for this summer’s work, and are planning to take another look at the robustness of the logic in the face of network timeouts and the like.

1 Like

Thanks both, looks like it’s on the radar for being addressed.