Bug - Refreshing the jobs statuses seems to use an "unknown" value for failed_at

Summary

The log file reports that while the computations are happening on the cluster, it returns an "unknown" value for the failed_at and therefore tries to use this value to fill the DB, which doesn't work, preventing failure while the computations are still ongoing.

Steps to reproduce

Create a job and look at the application log file.

What is the current bug behavior?

The refresh jobs command seems to retrieve an Unknown value for failed_at.

What is the expected correct behavior?

The refresh jobs command should retrieve no value for failed_at and therefore should not cause an error.

Relevant screenshots and/or logs

[2021-11-05 13:18:01] local.ERROR: SQLSTATE[22007]: Invalid datetime format: 1292 Incorrect datetime value: 'Unknown' for column 'failed_at' at row 1 (SQL: update `cluster_jobs` set `failed_at` = Unknown where `id` = 29) {"exception":"[object] (Illuminate\\Database\\QueryException(code: 22007): SQLSTATE[22007]: Invalid datetime format: 1292 Incorrect datetime value: 'Unknown' for column 'failed_at' at row 1 (SQL: update `cluster_jobs` set `failed_at` = Unknown where `id` = 29) at /var/vhosts/sib.swiss/oncobench-dev/htdocs/app/vendor/laravel/framework/src/Illuminate/Database/Connection.php:703)