CCAR-P Advanced Agentic Architecture Practice Question
An agentic system is designed to run long-lived 'background' tasks that may take several days to complete. How should the architect manage the agent's state to ensure it can resume correctly after a system reboot?
⚠ Common exam trap
Candidates often assume the agent's session memory is sufficient. They ignore the reality that long-lived background tasks require persistent storage to survive system restarts, crashes, or scaling events.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Externalize state and history to a persistent store
Long-lived agents cannot rely on in-memory state. To ensure durability, the entire conversation history and any relevant internal state (like current goal or pending sub-tasks) must be externalized to a persistent database. This allows any worker instance to reconstruct the agent's 'mind' and continue the task from the last successful turn.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Pass the entire history in every API call
Why it's wrong here
While necessary for the model to have context, simply passing history in the API call doesn't solve the persistence problem. If the orchestrator crashes, the history in memory is lost. You need a way to store that history externally so it can be retrieved and passed to the API again after a reboot.
- ✗
Use a very long 'max_tokens' setting
Why it's wrong here
The 'max_tokens' parameter controls the length of a single model response, not the duration of an agentic session. It provides no protection against system failures or reboots and has no impact on the persistence of the conversation history or the agent's ability to resume after an interruption.
- ✓
Externalize state and history to a persistent store
Why this is correct
By saving the messages array and any orchestrator-level metadata (like step progress or tool IDs) to a database, the system becomes resilient. After a reboot, the orchestrator can reload the state, identify the last completed action, and resume the loop without losing progress or context.
- ✗
Rely on the model's internal memory
Why it's wrong here
LLMs are stateless; they do not have internal memory that persists between separate API calls or system restarts. Every request is independent. Any 'memory' must be provided explicitly in the prompt or messages array for each turn, making external persistence a requirement for long-lived agentic tasks.
About these practice questions
Courseiva writes every CCAR-P question from scratch — 262 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCAR-P practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCAR-P exam.