Improve jobs resistence to temporary storage failures

Summary:

• Use more repetitions (most jobs will use $tries=24)
• Unify backoff values across jobs
• Use longer backoff intervals after an error
• Remove redundant configuration/code

Reviewers: #hkccp_developers, mollekopf

Reviewed By: #hkccp_developers, mollekopf

Subscribers: mollekopf, #hkccp_developers

Differential Revision: https://git.kolab.org/D5124