Model
Retrying tasks, not jobs
Job-level success is not something you can buy from reliable parts; it comes from retrying at the task level. This works only if task failures are independent.
- Tasks
- tasks in the job
- Task failure
- failure probability of one task attempt
- Attempts
- attempts allowed per task
How it behaves
One input on the horizontal axis, the others held fixed.
no retry (k = 1)k = 2k = 3
Data table
| Tasks n | series | P(job succeeds) |
|---|---|---|
| 10 | no retry (k = 1) | 0.99 |
| 11.9 | no retry (k = 1) | 0.988 |
| 14.1 | no retry (k = 1) | 0.986 |
| 16.8 | no retry (k = 1) | 0.983 |
| 20 | no retry (k = 1) | 0.98 |
| 23.7 | no retry (k = 1) | 0.977 |
| 28.2 | no retry (k = 1) | 0.972 |
| 33.5 | no retry (k = 1) | 0.967 |
| 39.8 | no retry (k = 1) | 0.961 |
| 47.3 | no retry (k = 1) | 0.954 |
| 56.2 | no retry (k = 1) | 0.945 |
| 66.8 | no retry (k = 1) | 0.935 |
| 79.4 | no retry (k = 1) | 0.924 |
| 94.4 | no retry (k = 1) | 0.91 |
| 112 | no retry (k = 1) | 0.894 |
| 133 | no retry (k = 1) | 0.875 |
| 158 | no retry (k = 1) | 0.853 |
| 188 | no retry (k = 1) | 0.828 |
| 224 | no retry (k = 1) | 0.799 |
| 266 | no retry (k = 1) | 0.766 |
| 316 | no retry (k = 1) | 0.729 |
| 376 | no retry (k = 1) | 0.687 |
| 447 | no retry (k = 1) | 0.64 |
| 531 | no retry (k = 1) | 0.588 |
| 631 | no retry (k = 1) | 0.532 |
| 750 | no retry (k = 1) | 0.472 |
| 891 | no retry (k = 1) | 0.41 |
| 1.06K | no retry (k = 1) | 0.347 |
| 1.26K | no retry (k = 1) | 0.284 |
| 1.5K | no retry (k = 1) | 0.224 |
| 1.78K | no retry (k = 1) | 0.169 |
| 2.11K | no retry (k = 1) | 0.121 |
| 2.51K | no retry (k = 1) | 0.081 |
| 2.99K | no retry (k = 1) | 0.05 |
| 3.55K | no retry (k = 1) | 0.029 |
| 4.22K | no retry (k = 1) | 0.015 |
| 5.01K | no retry (k = 1) | 0.007 |
| 5.96K | no retry (k = 1) | 0.003 |
| 7.08K | no retry (k = 1) | 0.001 |
| 8.41K | no retry (k = 1) | 0 |
| 10K | no retry (k = 1) | 0 |
| 11.9K | no retry (k = 1) | 0 |
| 14.1K | no retry (k = 1) | 0 |
| 16.8K | no retry (k = 1) | 0 |
| 20K | no retry (k = 1) | 0 |
| 23.7K | no retry (k = 1) | 0 |
| 28.2K | no retry (k = 1) | 0 |
| 33.5K | no retry (k = 1) | 0 |
| 39.8K | no retry (k = 1) | 0 |
| 47.3K | no retry (k = 1) | 0 |
| 56.2K | no retry (k = 1) | 0 |
| 66.8K | no retry (k = 1) | 0 |
| 79.4K | no retry (k = 1) | 0 |
| 94.4K | no retry (k = 1) | 0 |
| 112K | no retry (k = 1) | 0 |
| 133K | no retry (k = 1) | 0 |
| 158K | no retry (k = 1) | 0 |
| 188K | no retry (k = 1) | 0 |
| 224K | no retry (k = 1) | 0 |
| 266K | no retry (k = 1) | 0 |
| 316K | no retry (k = 1) | 0 |
| 376K | no retry (k = 1) | 0 |
| 447K | no retry (k = 1) | 0 |
| 531K | no retry (k = 1) | 0 |
| 631K | no retry (k = 1) | 0 |
| 750K | no retry (k = 1) | 0 |
| 891K | no retry (k = 1) | 0 |
| 1.06M | no retry (k = 1) | 0 |
| 1.26M | no retry (k = 1) | 0 |
| 1.5M | no retry (k = 1) | 0 |
| 1.78M | no retry (k = 1) | 0 |
| 2.11M | no retry (k = 1) | 0 |
| 2.51M | no retry (k = 1) | 0 |
| 2.99M | no retry (k = 1) | 0 |
| 3.55M | no retry (k = 1) | 0 |
| 4.22M | no retry (k = 1) | 0 |
| 5.01M | no retry (k = 1) | 0 |
| 5.96M | no retry (k = 1) | 0 |
| 7.08M | no retry (k = 1) | 0 |
| 8.41M | no retry (k = 1) | 0 |
| 10M | no retry (k = 1) | 0 |
| 10 | k = 2 | 1 |
| 11.9 | k = 2 | 1 |
| 14.1 | k = 2 | 1 |
| 16.8 | k = 2 | 1 |
| 20 | k = 2 | 1 |
| 23.7 | k = 2 | 1 |
| 28.2 | k = 2 | 1 |
| 33.5 | k = 2 | 1 |
| 39.8 | k = 2 | 1 |
| 47.3 | k = 2 | 1 |
| 56.2 | k = 2 | 1 |
| 66.8 | k = 2 | 1 |
| 79.4 | k = 2 | 1 |
| 94.4 | k = 2 | 1 |
| 112 | k = 2 | 1 |
| 133 | k = 2 | 1 |
| 158 | k = 2 | 1 |
| 188 | k = 2 | 1 |
| 224 | k = 2 | 1 |
| 266 | k = 2 | 1 |
| 316 | k = 2 | 1 |
| 376 | k = 2 | 1 |
| 447 | k = 2 | 1 |
| 531 | k = 2 | 0.999 |
| 631 | k = 2 | 0.999 |
| 750 | k = 2 | 0.999 |
| 891 | k = 2 | 0.999 |
| 1.06K | k = 2 | 0.999 |
| 1.26K | k = 2 | 0.999 |
| 1.5K | k = 2 | 0.999 |
| 1.78K | k = 2 | 0.998 |
| 2.11K | k = 2 | 0.998 |
| 2.51K | k = 2 | 0.997 |
| 2.99K | k = 2 | 0.997 |
| 3.55K | k = 2 | 0.996 |
| 4.22K | k = 2 | 0.996 |
| 5.01K | k = 2 | 0.995 |
| 5.96K | k = 2 | 0.994 |
| 7.08K | k = 2 | 0.993 |
| 8.41K | k = 2 | 0.992 |
| 10K | k = 2 | 0.99 |
| 11.9K | k = 2 | 0.988 |
| 14.1K | k = 2 | 0.986 |
| 16.8K | k = 2 | 0.983 |
| 20K | k = 2 | 0.98 |
| 23.7K | k = 2 | 0.977 |
| 28.2K | k = 2 | 0.972 |
| 33.5K | k = 2 | 0.967 |
| 39.8K | k = 2 | 0.961 |
| 47.3K | k = 2 | 0.954 |
| 56.2K | k = 2 | 0.945 |
| 66.8K | k = 2 | 0.935 |
| 79.4K | k = 2 | 0.924 |
| 94.4K | k = 2 | 0.91 |
| 112K | k = 2 | 0.894 |
| 133K | k = 2 | 0.875 |
| 158K | k = 2 | 0.853 |
| 188K | k = 2 | 0.828 |
| 224K | k = 2 | 0.799 |
| 266K | k = 2 | 0.766 |
| 316K | k = 2 | 0.729 |
| 376K | k = 2 | 0.687 |
| 447K | k = 2 | 0.64 |
| 531K | k = 2 | 0.588 |
| 631K | k = 2 | 0.532 |
| 750K | k = 2 | 0.472 |
| 891K | k = 2 | 0.41 |
| 1.06M | k = 2 | 0.347 |
| 1.26M | k = 2 | 0.284 |
| 1.5M | k = 2 | 0.224 |
| 1.78M | k = 2 | 0.169 |
| 2.11M | k = 2 | 0.121 |
| 2.51M | k = 2 | 0.081 |
| 2.99M | k = 2 | 0.051 |
| 3.55M | k = 2 | 0.029 |
| 4.22M | k = 2 | 0.015 |
| 5.01M | k = 2 | 0.007 |
| 5.96M | k = 2 | 0.003 |
| 7.08M | k = 2 | 0.001 |
| 8.41M | k = 2 | 0 |
| 10M | k = 2 | 0 |
| 10 | k = 3 | 1 |
| 11.9 | k = 3 | 1 |
| 14.1 | k = 3 | 1 |
| 16.8 | k = 3 | 1 |
| 20 | k = 3 | 1 |
| 23.7 | k = 3 | 1 |
| 28.2 | k = 3 | 1 |
| 33.5 | k = 3 | 1 |
| 39.8 | k = 3 | 1 |
| 47.3 | k = 3 | 1 |
| 56.2 | k = 3 | 1 |
| 66.8 | k = 3 | 1 |
| 79.4 | k = 3 | 1 |
| 94.4 | k = 3 | 1 |
| 112 | k = 3 | 1 |
| 133 | k = 3 | 1 |
| 158 | k = 3 | 1 |
| 188 | k = 3 | 1 |
| 224 | k = 3 | 1 |
| 266 | k = 3 | 1 |
| 316 | k = 3 | 1 |
| 376 | k = 3 | 1 |
| 447 | k = 3 | 1 |
| 531 | k = 3 | 1 |
| 631 | k = 3 | 1 |
| 750 | k = 3 | 1 |
| 891 | k = 3 | 1 |
| 1.06K | k = 3 | 1 |
| 1.26K | k = 3 | 1 |
| 1.5K | k = 3 | 1 |
| 1.78K | k = 3 | 1 |
| 2.11K | k = 3 | 1 |
| 2.51K | k = 3 | 1 |
| 2.99K | k = 3 | 1 |
| 3.55K | k = 3 | 1 |
| 4.22K | k = 3 | 1 |
| 5.01K | k = 3 | 1 |
| 5.96K | k = 3 | 1 |
| 7.08K | k = 3 | 1 |
| 8.41K | k = 3 | 1 |
| 10K | k = 3 | 1 |
| 11.9K | k = 3 | 1 |
| 14.1K | k = 3 | 1 |
| 16.8K | k = 3 | 1 |
| 20K | k = 3 | 1 |
| 23.7K | k = 3 | 1 |
| 28.2K | k = 3 | 1 |
| 33.5K | k = 3 | 1 |
| 39.8K | k = 3 | 1 |
| 47.3K | k = 3 | 1 |
| 56.2K | k = 3 | 1 |
| 66.8K | k = 3 | 1 |
| 79.4K | k = 3 | 1 |
| 94.4K | k = 3 | 1 |
| 112K | k = 3 | 1 |
| 133K | k = 3 | 1 |
| 158K | k = 3 | 1 |
| 188K | k = 3 | 1 |
| 224K | k = 3 | 1 |
| 266K | k = 3 | 1 |
| 316K | k = 3 | 1 |
| 376K | k = 3 | 1 |
| 447K | k = 3 | 1 |
| 531K | k = 3 | 0.999 |
| 631K | k = 3 | 0.999 |
| 750K | k = 3 | 0.999 |
| 891K | k = 3 | 0.999 |
| 1.06M | k = 3 | 0.999 |
| 1.26M | k = 3 | 0.999 |
| 1.5M | k = 3 | 0.999 |
| 1.78M | k = 3 | 0.998 |
| 2.11M | k = 3 | 0.998 |
| 2.51M | k = 3 | 0.997 |
| 2.99M | k = 3 | 0.997 |
| 3.55M | k = 3 | 0.996 |
| 4.22M | k = 3 | 0.996 |
| 5.01M | k = 3 | 0.995 |
| 5.96M | k = 3 | 0.994 |
| 7.08M | k = 3 | 0.993 |
| 8.41M | k = 3 | 0.992 |
| 10M | k = 3 | 0.99 |
Worked examples, recomputed
Every number the mechanisms quote for this relationship is recomputed from it, so none of them disagrees with it.
Where it appears
Each mechanism states it in its own names. The general symbol each name stands for is shown with it.
- Fault Tolerance & Delivery Semantics
What is the cheapest thing that can be safely done twice?