All Cisco exams

Free 350-901 practice test: 30 questions with answers

30 original questions for Designing, Deploying and Managing Network Automation Systems (AUTOCOR), drawn from every topic area. Answer each one to see whether you got it right and why. No sign-up.

Free questions
30
Full bank
100 questions

350-901 exam guide: topics, cost and format

All 30 350-901 questions with answers

The same questions as the test above, grouped by exam topic. Open any question to see the answer and explanation.

Network Programmability and Automation Design (10 questions)

  1. Question 1

    A network team wants to change VLANs on many switches with a single operation that is validated before it is applied. Which NETCONF capability supports validation before commit?

    1. AThe validate operation on the candidate configuration
    2. BThe show running-config command on each switch
    3. CA manual SSH session to each device, editing the VLAN lines by hand
    4. DAn SNMP trap listener receiving status alerts from devices
    Show answer and explanation
    Correct answer: A

    The validate operation checks the candidate configuration before it is committed. Showing running configuration, manual sessions and trap listeners do not validate changes.

  2. Question 2

    An application needs to subscribe to changes in interface state on a router and receive updates when they occur. Which mechanism fits best?

    1. APolling with SNMP GET every second for all objects
    2. BReading the device's log file each day
    3. CEmailing the operator when a change occurs
    4. DNETCONF notifications or YANG-push subscriptions
    Show answer and explanation
    Correct answer: D

    NETCONF notifications and YANG-push deliver changes as they occur. Frequent polling is inefficient, daily log reads are delayed, and email is manual.

  3. Question 3

    Which HTTP method is used in a REST API to replace an existing resource entirely?

    1. ADELETE
    2. BOPTIONS
    3. CPUT
    4. DGET
    Show answer and explanation
    Correct answer: C

    PUT replaces the resource with the supplied representation. GET reads, DELETE removes, and OPTIONS describes allowed methods.

  4. Question 4

    Which HTTP method should be used to change only part of an existing resource, such as a single interface description?

    1. ATRACE
    2. BPATCH
    3. CPUT only
    4. DHEAD
    Show answer and explanation
    Correct answer: B

    PATCH updates part of a resource. PUT replaces the entire resource, HEAD returns headers only, and TRACE is a diagnostic method.

  5. Question 5

    A network automation system must represent devices as a single source of truth with their interfaces, VLANs, and policies. Which approach supports this design?

    1. AKeeping configurations in individual engineers' paper notebooks and personal spreadsheets, updated by hand
    2. BUsing whichever configuration the device currently has in running memory
    3. CStoring device details in a shared chat thread with IP addresses and VLAN lists
    4. DA central inventory and source-of-truth database that automation reads from and writes to
    Show answer and explanation
    Correct answer: D

    A central source of truth keeps intended state consistent and drives automation. Notebooks, chat threads and current device state alone do not provide one authoritative model.

  6. Question 6

    A controller exposes a northbound API for applications and a southbound interface to devices. What is the role of the northbound API?

    1. ASending raw packets directly to end users' hosts, bypassing the service policy
    2. BManaging physical cabling, patch panel labeling, and rack power distribution
    3. CReplacing the operating system image of each device with a firmware build during a maintenance window
    4. DAllowing applications and automation to request network services from the controller
    Show answer and explanation
    Correct answer: D

    Northbound APIs let applications and automation use the controller. Packet delivery, cabling and OS replacement are not northbound functions.

  7. Question 7

    Which software design principle keeps automation modules independent so that a change to one module does not break others?

    1. ACopying the same function into each script separately by hand
    2. BModularity with clear interfaces between components
    3. CPutting the code in one file with global variables
    4. DAvoiding written documentation of interfaces
    Show answer and explanation
    Correct answer: B

    Modular design with clear interfaces isolates changes and eases testing. Single files with global state, copied functions and undocumented interfaces increase coupling.

  8. Question 8

    A team wants to test a Python automation module without touching real devices. Which technique is most appropriate?

    1. ATesting by reading the code aloud during a team review meeting with colleagues
    2. BUsing mock objects or simulated device responses in unit tests
    3. CRunning the module against live production routers in the core network
    4. DSkipping unit tests because automation scripts are simple and low risk
    Show answer and explanation
    Correct answer: B

    Mocks and simulated responses allow tests without impacting devices. Production testing, skipping and informal reading do not verify behavior safely.

  9. Question 9

    An automation workflow must wait for a device to finish a long-running operation before continuing. Which technique fits best?

    1. ASleeping for an arbitrary hour before continuing
    2. BContinuing right away without checking status
    3. CPolling the operation status with a timeout and backoff
    4. DSending the job by email and waiting for a reply
    Show answer and explanation
    Correct answer: C

    Polling with timeout and backoff checks progress efficiently and fails cleanly if the operation stalls. Fixed long sleeps, unchecked continuation and email waits are unreliable.

  10. Question 10

    A workflow creates a resource in a controller, and the request may be retried after a timeout. Which design avoids duplicate resources?

    1. ACreating a new resource with a randomly generated name every time
    2. BSending the same create request blindly and repeatedly without checking whether the resource exists
    3. CDeleting existing resources in the controller before each retry attempt
    4. DUsing idempotent request design with unique identifiers or check-then-create logic
    Show answer and explanation
    Correct answer: D

    Idempotent requests and check-then-create logic prevent duplicates on retry. Random names, repeated blind requests and deleting everything create new problems.

Network Automation Deployment and Operations (10 questions)

  1. Question 11

    Which two items should be documented for a network automation runbook? (Choose two.)

    1. ARollback steps and escalation contacts
    2. BThe office coffee menu
    3. CThe developer's favorite editor
    4. DThe color scheme of the diagrams
    5. EPrerequisites, inputs, and expected outputs
    Show answer and explanation
    Correct answer: A, E

    Prerequisites, inputs, outputs and rollback steps with contacts make runbooks usable under pressure. Personal and cosmetic details are not operational content.

  2. Question 12

    An automation platform runs in containers that must be deployed on several hosts and restarted automatically if they fail. Which orchestration approach fits best?

    1. AKubernetes or a container orchestrator with health probes and restart policies
    2. BRunning containers on one developer laptop with local Docker Desktop
    3. CStarting each container by hand with docker run on every host and checking status over SSH
    4. DDisabling restart policies to keep day-to-day operations simpler
    Show answer and explanation
    Correct answer: A

    Container orchestrators manage placement, health checks and restarts across hosts. Manual starts, laptops and disabled restarts do not meet availability needs.

  3. Question 13

    A CI pipeline runs automated tests on every commit to an automation repository. Which stage should run first to catch errors quickly?

    1. AStatic analysis and syntax checks
    2. BArchiving a repository backup
    3. CProduction deployment rollout
    4. DManual approval from the CEO
    Show answer and explanation
    Correct answer: A

    Static analysis and syntax checks are fast and catch many errors before deeper testing. Production deployment and archiving should come after validation, and executive approval is not a test stage.

  4. Question 14

    An automation job must run every night and report failures to the operations team. Which design achieves this reliably?

    1. AA job that runs without logging, leaving failures unnoticed until morning
    2. BA scheduled job with logging, alerts on failure, and retry handling
    3. CA job that runs once and is not checked again after its run
    4. DA manual reminder email each month asking an engineer to check the job output
    Show answer and explanation
    Correct answer: B

    Scheduled jobs with logs and failure alerts provide reliable operations. One-time runs, monthly reminders and unlogged jobs cannot show whether automation worked.

  5. Question 15

    A team must store API credentials for automation running in a CI pipeline. Which practice is recommended?

    1. ACommitting credentials to the repository's configuration file, where git history keeps them
    2. BPasting credentials into the pipeline logs, where they stay visible
    3. CStoring credentials in the pipeline's secret store and injecting them at run time
    4. DSharing credentials in a team chat channel with a message to the group
    Show answer and explanation
    Correct answer: C

    A pipeline secret store injects credentials at run time without exposing them in code or logs. Repository commits, logs and chat messages expose secrets.

  6. Question 16

    Which Cisco product provides telemetry collection and analytics for network health across a campus or data center, and can be used to drive automation decisions?

    1. ACisco Webex Meetings hosted video conferencing for remote teams
    2. BCisco Duo multifactor authentication for remote access
    3. CCisco Network Insights or Catalyst Center assurance
    4. DCisco Umbrella DNS-layer security for filtering risky sites
    Show answer and explanation
    Correct answer: C

    Assurance and insights products analyze network telemetry to assess health and inform automation. Webex, Duo and Umbrella do not analyze network health telemetry.

  7. Question 17

    A network operator wants to know whether a recent automated change improved packet loss. Which data is most useful?

    1. AThe change ticket's color label set by the ticketing tool for each request
    2. BThe operator's favorite dashboard theme picked for the team's review
    3. CPacket loss metrics collected before and after the change for the affected links
    4. DThe number of automation scripts in the repository and their line counts
    Show answer and explanation
    Correct answer: C

    Before-and-after metrics show whether the change improved the outcome. Repository counts, labels and themes do not measure network performance.

  8. Question 18

    A controller must receive events from many devices, and the events must be processed reliably even if the controller is briefly busy. Which pattern fits best?

    1. AUsing a message queue or event bus between devices and processing services
    2. BWriting events to a shared spreadsheet that each device updates by hand during the shift
    3. CSending events directly to a single thread with no buffer in the main polling loop
    4. DDropping events when the controller is busy and silently discarding the event data
    Show answer and explanation
    Correct answer: A

    A message queue or event bus buffers events and decouples producers from consumers. Single-thread direct delivery, dropping events and spreadsheets are unreliable.

  9. Question 19

    A network team wants a dashboard that shows automation job success rates over time. Which practice supports this?

    1. ACounting the jobs manually once a year by walking through the scheduler's logs
    2. BExporting job results as structured metrics to a monitoring system
    3. CKeeping results in each engineer's notebook with handwritten status notes
    4. DIgnoring job failures and continuing the nightly schedule without review
    Show answer and explanation
    Correct answer: B

    Structured metrics in a monitoring system support trend dashboards. Yearly counts, personal notes and ignored failures do not provide timely visibility.

  10. Question 20

    An automation script must avoid overloading a device with too many API calls at once. Which technique helps most?

    1. ARate limiting and concurrency limits per device
    2. BRepeating each failed request ten times in quick succession
    3. CSending as many parallel requests as possible at once
    4. DDisabling API timeouts for device session connections
    Show answer and explanation
    Correct answer: A

    Rate and concurrency limits protect devices from overload. Maximum parallelism, repeated requests and disabled timeouts cause overload and hangs.

Network Automation Fundamentals (10 questions)

  1. Question 21

    An engineer wants to retrieve the interface configuration from a Cisco device over an API using a standard HTTP method and a JSON-friendly response. Which API style is most appropriate?

    1. ATelnet sessions with copy and paste
    2. BFTP transfers of configuration files only
    3. CA paper change request form
    4. DRESTCONF or REST API with GET requests
    Show answer and explanation
    Correct answer: D

    RESTCONF and REST APIs use HTTP GET to read configuration and return structured data such as JSON. Paper forms, Telnet and FTP do not provide structured API access.

  2. Question 22

    Which data modeling language is used by NETCONF and RESTCONF to describe the structure of configuration and state data?

    1. AYANG
    2. BHTML
    3. CCSS
    4. DSQL only
    Show answer and explanation
    Correct answer: A

    YANG models define the structure and constraints of network data for NETCONF and RESTCONF. HTML and CSS are web languages, and SQL is a database query language.

  3. Question 23

    A script must change configuration transactionally so that either all changes apply or none do. Which protocol provides this with candidate and commit operations?

    1. ASNMP GET
    2. BICMP
    3. CNETCONF
    4. DDHCP
    Show answer and explanation
    Correct answer: C

    NETCONF supports candidate configuration and commit operations, which allow transactional changes. SNMP GET reads data, ICMP tests reachability, and DHCP assigns addresses.

  4. Question 24

    Which HTTP status code indicates that the API request was accepted but the client did not supply valid credentials?

    1. A500 Internal Server Error
    2. B401 Unauthorized
    3. C200 OK
    4. D404 Not Found
    Show answer and explanation
    Correct answer: B

    401 means authentication is missing or invalid. 200 indicates success, 500 indicates a server error, and 404 indicates the resource was not found.

  5. Question 25

    Which encoding format is commonly used in REST API responses and is easy for scripts to parse?

    1. APrinted PDFs
    2. BJSON
    3. CBitmap images
    4. DPlain screenshots
    Show answer and explanation
    Correct answer: B

    JSON is a lightweight structured format that scripts parse easily. Images, screenshots and PDFs are not structured data formats for APIs.

  6. Question 26

    A network team wants to store its automation scripts with a full history of changes and the ability to roll back. Which tool should be used?

    1. AA shared folder with files named final_v2
    2. BGit version control
    3. CEmail attachments
    4. DA printed copy of the code
    Show answer and explanation
    Correct answer: B

    Git records history, supports branching and rollback, and enables peer review. Shared folders with ad hoc names, email and printouts do not provide reliable history.

  7. Question 27

    An automation pipeline must test changes on a virtual lab before they reach production devices. Which practice is recommended?

    1. ARunning validation tests in a lab or simulated environment as part of the CI/CD pipeline
    2. BSkipping the test stage for small configuration changes because they are judged low risk by peers
    3. CTesting changes directly on live production routers during peak business hours, without a staging step
    4. DWaiting until a major production outage occurs, then running validation tests on the affected routers
    Show answer and explanation
    Correct answer: A

    Lab validation in the pipeline catches errors before production. Production testing, skipped tests and after-outage testing increase risk.

  8. Question 28

    Which Python practice keeps network credentials out of automation scripts?

    1. APrinting the plaintext password to the log file during debugging
    2. BStoring the password in a commented-out header line at the top of the script
    3. CReading secrets from environment variables or a secure secrets store at run time
    4. DHardcoding the device password as a string literal in the connection settings of the script file
    Show answer and explanation
    Correct answer: C

    Environment variables or secret stores keep credentials out of code and version control. Hardcoded, logged and commented passwords expose them.

  9. Question 29

    A script uses an API that returns paginated results. What must the script do to collect all results?

    1. AFollow the next-page links or offsets until no more pages are returned
    2. BConvert the returned JSON results into rendered image files
    3. CRead the first page of results and then end the loop early
    4. DRepeat the initial request indefinitely with the same query parameters
    Show answer and explanation
    Correct answer: A

    Following pagination links or offsets collects the complete data set. Stopping early or repeating the first request returns incomplete or duplicate data.

  10. Question 30

    An automation script must be idempotent. Which behavior describes idempotence?

    1. AEach run creates a new object with a randomly generated name instead of updating the existing one
    2. BThe script aborts with an error if the configuration already exists
    3. CRunning the script multiple times produces the same end state as running it once
    4. DThe script deletes the existing configuration on every run before applying it
    Show answer and explanation
    Correct answer: C

    Idempotent operations converge to the same state no matter how many times they run. The other behaviors create duplicates, fail unnecessarily or remove configuration.