SAP-C02 Practice Question 49: Cron Fleet Modernization
AWS Certified Solutions Architect – Professional (SAP-C02) · Test 5 · question 9 of 10
A retailer runs about 400 scheduled jobs from crontab on two "utility" EC2 instances in a single Availability Zone. The jobs start Amazon ECS tasks, invoke AWS Lambda functions, publish to Amazon SNS, and call other AWS APIs directly.
The arrangement has become a liability. When one instance was terminated by a failed patch run, 200 jobs silently stopped for a day before anyone noticed. A failed job never retries. Several jobs must fire at a local wall-clock time in Europe and North America, and twice a year the daylight saving change has caused them to run an hour late. Nobody wants to keep patching the instances.
The team wants the schedules moved to a managed service with per-schedule retry behaviour, native time-zone support, and no servers to maintain. What should the solutions architect recommend?
Show answer and explanation
Answer:
EventBridge Scheduler is a serverless scheduler built for exactly this migration. It creates, runs, and manages tasks from one managed service with no infrastructure to provision, which retires the utility instances and the patching that killed them. Its universal target parameter can invoke more than 270 AWS services and over 6,000 API operations, so the ECS, Lambda, SNS, and other API calls become targets directly rather than needing glue code. Schedules are defined with cron and rate expressions for recurring patterns or as one-time invocations, and schedules carry a time zone, which is what makes the European and North American wall-clock jobs survive daylight saving changes. And EventBridge Scheduler lets you set the number of retries and the maximum retention time for a failed task per schedule, replacing the silent failures of cron.
Option A makes the fleet more resilient but keeps every problem the team actually listed: servers to patch, no retry semantics, no time-zone handling, and now a distributed lock to write and debug.
Option C is the closest wrong answer. EventBridge rules do schedule invocations, but their cron expressions are evaluated in UTC with no time-zone support, so the daylight saving problem survives, and each of the 400 jobs gains a Lambda function to write, deploy, and maintain purely as a relay.
Option D is the wrong tool. State Manager keeps managed nodes in a defined configuration state; it is node-centric, so it neither removes the instances nor provides a natural way to invoke arbitrary AWS APIs on a schedule.
Reference: What is Amazon EventBridge Scheduler?