Closed jaakkone closed 7 years ago
Yes that happened because of the heavy workload that was going to the frontend server to check for the status of Jani's jobs. This needs to be fixed after the proof of concept to allow large submissions of jobs without freezing the web application server.
We are now doing less requests to the server to check for the job status. In the following days this will be further optimised to perform less requests when checking for the job id associated to each process.
We now only perform a single request for each pipeline to get the job ids and then single requests in a more wide interval to check for the job status.
The web app of our platform (https://193.166.24.85/app) responded poorly (long response time, missing dialogs, e.g. File 1 and File 2 when adding a new strain, could not view Reports) while Jani was doing heavy calculation (INNUca for >100 strains, 19 Oct 2017). Should there be some resources left (priorities) on the calculation servers to answer calls from the web app?