33. Background Job Processing

Date: 2024-04-13

Status

Accepted

Context

The need for efficient background job processing arises from requirements such as handling big data operations, managing long-running tasks, and automating push notifications, report generation, and email/newsletter dispatches. Additionally, caching strategies like invalidation, hot-cache management, and pre-caching require reliable scheduling and execution mechanisms.

Decision

To address these needs, the following options have been considered for implementing background jobs:

• Kubernetes Jobs:
  • CronJob: For recurring tasks based on a time-based schedule. More about Kubernetes CronJobs
  • Job: For one-off tasks that need to be run to completion. More about Kubernetes Jobs

• Temporal: A workflow orchestration platform ideal for complex workflows and handling retries, failures, and state pass-through between tasks. Learn more about Temporal

• Standard Cron: Utilizing the operating system’s cron for straightforward, time-based job scheduling.

• For Golang Applications:
  • Golang Cron Library: A library suited for scheduling recurring tasks within Go applications. Visit Golang Cron Library
  • Go Ticker and Timer: Utilizing native Go features for fine-grained time-based operations.
• Message Queueing (MQ):
  • Apache Pulsar: For decoupling job requests from job execution, supporting event-driven architectures and ensuring scalable message processing. Explore Apache Pulsar

Consequences

Technology	Benefits	Challenges
Standard Cron	Simple and reliable for scheduled tasks.	Limited to single-node setups, lacks fault tolerance.
Kubernetes CronJob	Scales well with Kubernetes, handles cluster-wide tasks.	More complex setup, requires Kubernetes management.
Temporal	Provides robust handling of complex workflows, retries, and failures.	Higher learning curve, more resource-intensive.
Go Ticker/Timer	High precision, integrated directly in Go applications.	Requires custom setup for distributed tasks, more manual management of task state.