If you are running SQL Server jobs with numerous steps there can be failures due to network glitches or deadlocks. In such a scenario the DBA will typically restart the job from the failed step. However, there is an option in SQL Server Agent to automatically retry the job, which is under utilized by many DBAs.
For each step of the job, in the advanced tab, you have an option of retry attempts and retry intervals.
In case of network glitches you should allow at least 2-3 minutes to recover. So taking that into consideration, you can use set the retry interval to suit your environment.
This retry option will reflect in the Job history as well :