Pick a focused question that fits your time, stack, and interview goal.
How much time do you have?
Show one-drill sessions you can finish now.
1225 drills fit a 20-minute session. Closest fits first.
Page 4 of 52
Finds synchronous I/O or CPU work that prevents the event loop from scheduling other tasks.
Combines asyncio debug evidence, task inspection, warnings, profiling, and capacity signals.
Defines fail-fast, all-results, and partial-success behavior for concurrent asyncio operations.
Places semaphore and admission boundaries around scarce work without retaining unlimited tasks.
Produces a pinned, reviewable Python artifact once and promotes the same image across environments.
Chooses atomic database expressions, row locks, and constraints from the invariant being protected.
Separates single scheduler ownership from task-level overlap and idempotency controls.
Connects pool limits to process topology, transaction duration, database capacity, and queueing evidence.
Uses expected versions and conditional writes so delayed or retried Python tasks cannot commit stale decisions.
Verifies commit callbacks without confusing callback registration with durable external delivery.
Overrides FastAPI boundaries for focused tests and restores the global override mapping reliably.
Trusts proxy-derived scheme, host, and client data only from known hops and configures externally visible paths correctly.
Uses plans as bounded, revisable execution aids while preserving evidence, policy, and application-owned state transitions.
Uses documented thread-safe scheduling APIs and keeps event-loop objects confined.
Uses a password-specific algorithm, measurable work settings, and gradual rehashing without blocking service capacity.
Reuses bounded redis-py pools for the application lifetime and closes async resources deliberately.
Coordinates readiness, signals, request drain, lifespan cleanup, and a bounded termination deadline.
Separates stale checkout protection, connection aging, pool waits, statement timeouts, and uncertain transactions.
Treats cancellation as cooperative control flow with reliable resource cleanup and propagation.
Uses FastAPI lifespan to create and close process-owned resources around request serving.
Keeps Celery messages small, safe to deserialize, and compatible across rolling deployments.
Aligns pytest-asyncio discovery, loop lifetime, and async fixture ownership without custom loop competition.
Defines one explicit SQLAlchemy session and transaction owner for each request, job, thread, or async task.
Uses deadline budgets, transient-only bounded retries, jitter, idempotency, and propagated cancellation without retry storms.