MCPA-LEVEL-1 practice question 3 of 15
A new upstream API Is being designed to offer an SLA of 500 ms median and 800 ms maximum (99th percentile) response time. The corresponding API…
Full set: 95 questionsThe upstream API requires a 99th percentile response time of 800 ms for a process that involves three sequential downstream calls. The provided downstream API has a 95th percentile response time of 1000 ms for a single call. This means that 5% of the time, a single downstream call will take longer than the entire time budget for the upstream API, guaranteeing an SLA violation even before the other two calls are made. To enforce the 800 ms upstream SLA, a timeout would need to be set for each downstream call (e.g., ~260 ms). However, the downstream API's 80th percentile is 500 ms, meaning 20% of its calls take longer than 500 ms. A ~260 ms timeout would therefore fail a very high percentage of invocations, trading a latency problem for an availability crisis. The SLAs are fundamentally incompatible.
A. Set a timeout of 50 ms; this times out more invocations of that API but gives additional room for retries This is incorrect because a 50 ms timeout is below the downstream API's median response time of 100 ms, which would cause over 50% of calls to fail, making the service unusable.
B. Set a timeout of 100 ms; that leaves 400 ms for the other two downstream APIs to complete This is incorrect as it only considers the median SLA and ignores the stricter 99th percentile requirement. A 100 ms timeout (the median) would result in a ~50% failure rate.
D. Do not set a timeout; the Invocation of this API Is mandatory and so we must wait until it responds This is incorrect because not setting a timeout guarantees the upstream API will violate its 99th percentile SLA, as the downstream API's tail latency frequently exceeds the total allowed time. ---
1. MuleSoft Official Documentation, "Designing for Reliability": This documentation outlines principles for building resilient applications, including the use of timeouts. The core principle is that timeouts are meant to handle transient faults or unresponsive services, not to fix a fundamental mismatch in performance SLAs. Applying an aggressive timeout to a predictably slow dependency converts a latency problem into an availability problem, which is not a sound architectural solution. (Reference: MuleSoft Docs -> Anypoint Platform -> "Designing for Reliability" section).
2. MuleSoft Training, "Anypoint Platform Architecture: Application Networks": This course emphasizes the importance of defining and adhering to non-functional requirements (NFRs), including SLAs. A key architectural concept taught is that the SLAs of dependent services directly impact the achievable SLA of the consuming service. In this scenario, the composite SLA cannot be met due to the downstream dependency's characteristics. (Reference: Anypoint Platform Architecture: Application Networks course materials, Module on "Defining Non-Functional Requirements").
3. MuleSoft Whitepaper, "The Total Economic Impact™ Of MuleSoft Anypoint Platform": While a business-focused paper, it highlights the architectural principle of creating reusable, well-governed APIs with clear contracts (including SLAs). The scenario violates this principle by attempting to compose services with incompatible performance contracts, leading to a brittle and unreliable outcome. (Reference: Forrester Consulting, "The Total Economic Impact™ Of MuleSoft Anypoint Platform," May 2021, discussion on API governance and contracts).
Keep practicing
Take the free MCPA-LEVEL-1 practice test
Ten exam-style questions with answers and explanations, plus the exam facts and study guides.
More questions
Other MCPA-LEVEL-1 practice questions
- Question 2What is true about the technology architecture of Anypoint VPCs?
- Question 4An organization makes a strategic decision to move towards an IT operating model that emp…
- Question 5What is true about automating interactions with Anypoint Platform using tools such as Any…
- Question 6An organization is deploying their new implementation of the OrderStatus System API to mu…
- Question 7How can the application of a rate limiting API policy be accurately reflected in the RAML…
- Question 8The implementation of a Process API must change. What is a valid approach that minimizes …
