Reliability economics
Compare routing, cache, and retry economics
For engineering leaders testing how traffic routing and reliability controls change unit cost.
Open this preset in the workbenchDecision
What this route helps you decide
Expose whether cache gains outweigh retry cost and conditional model execution at the expected workload.
Inputs
- Probability that traffic reaches the modeled route
- Cache hit rate and cache-affected input volume
- Failure rate and maximum retry count
- Synthetic or user-entered model rates in USD
Outputs
- Submitted cost under conditional routing
- Retry-adjusted cost per successful task
- Typical user cost at the expected workload
- Modeled margin after reliability overhead
Worked example
Worked example
A routed call executes for 72% of tasks, sees a 42% cache hit rate, and retries failures no more than twice.
- Cost / successful task
- $0.0168
- Typical user / month
- $2.0174
- Modeled margin
- 76.9%
Assumptions
- Cache affects modeled input cost
- Failures are independent per attempt
- Execution probability is entered by the user
What this does not decide
- Does not benchmark latency
- Does not promise cache eligibility
- Does not infer provider terms
Reproduce the method
All values use synthetic or user-entered rates and deterministic formula version 2.0.0. Review the calculation methodology and source policy.