The MCD – Level 1 exam is closed-book: 60 multiple-choice questions in 120 minutes, with a 70% pass mark. That’s two minutes per question on average, which sounds generous until you hit a DataWeave transformation question that expects you to trace the output of a script in your head, with no editor to run it in. A meaningful share of the question set isn’t “define this term” — it’s “here’s a flow, here’s what’s in the payload at a given point, what happens next.” You’re expected to reason about Mule event structure (payload, attributes, variables) the way you would while actually debugging a flow in Anypoint Studio, not just recall a definition of it.
MuleSoft doesn’t publish numeric weightings per domain for this exam. What it does publish is the objective list broken into twelve named areas. Counting objectives per domain is the closest public proxy for where the exam concentrates, and it’s a reasonable one since each objective maps to roughly one testable concept:
Domain | Approx. objective count | Relative emphasis |
|---|---|---|
Processing records (For Each, Batch Job, watermarking, Object Store) | 10 | Heaviest |
Using connectors (Database, HTTP Request, Web Service Consumer, File/FTP, JMS) | 8 | Heavy |
Explaining application network basics | 7 | Heavy |
Handling errors | 7 | Heavy |
Transforming data with DataWeave | 7 | Heavy |
Deploying and managing APIs and integrations | 7 | Heavy |
Designing and consuming APIs (RAML) | 6 | Moderate |
Structuring Mule applications | 6 | Moderate |
Building API implementation interfaces (APIkit) | 5 | Moderate |
Accessing and modifying Mule events | 4 | Light |
Routing events (Choice, Scatter-Gather, Validation) | 3 | Light |
Debugging and troubleshooting Mule applications | 3 | Light |
Don’t read “light” as “skippable” — debugging and routing show up embedded inside scenario questions from other domains constantly, even though they’re small as standalone topics.
A few areas show up repeatedly in post-exam retrospectives from people who failed on a first attempt:
Data persistence rules across boundaries. Knowing exactly what survives — payload, attributes, variables — when a Mule event crosses a Flow Reference versus a connector boundary (an HTTP Request call, say) is tested precisely, not loosely. People who’ve only worked with simple linear flows tend to get this wrong under exam pressure.
On Error Continue vs. On Error Propagate. Candidates often “know” the difference conceptually but can’t predict, from a given flow diagram, which scope’s output actually reaches the caller and which one swallows the error. The exam tests the behavioral consequence, not the definition.
Calling Mule flows from a DataWeave script. This specific objective trips people up because it’s not covered as thoroughly in most self-paced walkthroughs as the more common “transform this JSON to XML” style of DataWeave question.
Batch Job vs. For Each — when each is actually appropriate. Both process collections, but the exam wants you to recognize which one fits a given scenario (asynchronous, chunked processing with aggregation vs. simple synchronous iteration), not just define them.
Automatic vs. manual watermarking. This is a narrow objective that’s easy to skip in review and shows up more often than its size in the objective list suggests.
Because the live exam is closed-book and scenario-heavy, drilling flashcard-style recall questions gets you only partway. The more useful habit is reading a short flow description or a snippet of DataWeave and predicting the output or behavior before checking the answer — that’s the actual skill being tested. Time yourself in blocks that mirror the pacing (roughly 2 minutes a question) rather than untimed review, since a chunk of the difficulty on exam day is cognitive load, not unfamiliarity with the material. When you miss a question, trace it back to which of the twelve domains it sits in — if misses cluster in “processing records” or “handling errors,” that’s where another pass pays off most, given how objective-dense those two domains are.
Mixing up query parameters and URI parameters in RAML definitions, especially when a scenario describes a resource and asks you to identify which one applies.
Misjudging what data is visible inside a Try scope’s error handler versus outside it.
Assuming DataWeave type coercion behaves the same for all source formats (JSON vs. XML vs. Java objects) when it doesn’t.
Treating APIkit-generated flows as a black box instead of understanding how a request actually gets routed through them.
Running out of time because the first 15–20 questions get over-analyzed; the exam rewards steady pacing more than perfectionism on early questions.
New to Mule 4, coming from general programming (Java, JavaScript, or similar): Budget 8–12 weeks of consistent, hands-on time in Anypoint Studio. Reading about DataWeave syntax isn’t a substitute for writing transformations against payloads — build small flows that touch each connector type (Database, HTTP, File, JMS) at least once.
Has related integration or API experience, new to MuleSoft specifically: 3–5 weeks is typical. Focus disproportionately on DataWeave syntax and Mule-specific error-handling scopes, since those are the parts that don’t transfer from general integration background.
Retaking after a failed attempt: 1–2 weeks, targeted at whichever domains your prior attempt’s weak areas fell into. A second attempt that repeats the same broad review tends to produce the same result.
Two things are worth tracking if you’re planning your timeline around this credential. First, as of mid-2026, Salesforce has been consolidating its certification naming across the MuleSoft line; this credential is now also referenced as “Salesforce Certified MuleSoft Developer” with the exam code Mule-Dev-201 alongside the long-standing MCD-Level-1 label. The content and objectives are the same credential — it’s a naming change, not a new exam — but if you’re registering, confirm which label the current registration portal uses so you book the right listing.
Second, the companion “MCD – Level 1 DELTA” exam — the shorter, 35-question path that let MCD – Integration and API Associate (Mule 3) holders upgrade without sitting the full exam — has been a feature of this certification track historically. If you hold the Mule 3 associate credential, verify at registration whether that abbreviated path is still being offered before you plan on it, since MuleSoft has trimmed legacy exam paths as its certification catalog has evolved.
Validity period has also been reported inconsistently across sources recently — historically two years from the pass date, with some current listings suggesting an extension. Confirm the validity term shown at the time you register rather than relying on either figure from memory.
Last updated on Oct, 14 2026
