cancel
Showing results for 
Search instead for 
Did you mean: 
Data Engineering
cancel
Showing results for 
Search instead for 
Did you mean: 

Select job cluster vs all purpose cluster

greyamber
New Contributor II

I have workflow and need to run at every 1 minute interval, it is rest api call, should I go for all purpose cluster or job cluster to meet the SLA. We need to get the as soon as it is available. 

6 REPLIES 6

Kaniz
Community Manager
Community Manager

Hi @greyamber

Cluster Purpose:

SLA Requirements:

  • A job cluster might be more suitable if your workflow requires meeting a strict Service Level Agreement (SLA). Job clusters are dedicated to a specific task and can be optimized for performance.
  • All-purpose clusters, while versatile, may not provide the same level of predictability in meeting SLAs.

Resource Usage and Cost:

  • All-purpose clusters are more cost-effective for tasks like exploration and development, where resource sharing is beneficial.
  • Job clusters are more efficient for focused, time-sensitive tasks, as they release resources promptly after completion.

Trade-offs:

  • All-purpose clusters allow flexibility but may not guarantee immediate availability due to resource sharing.
  • Job clusters prioritize your specific job but come with higher resource costs.

Given your requirement for immediate availability, I recommend using a job cluster. It ensures that your REST API calls run promptly and efficiently, meeting your SLA. The cluster terminates once the processing is complete, minimizing resource usage and cost.

 

greyamber
New Contributor II

I’m bit confused. 
when I scheduled the job in job cluster. It took nearly 3 mins to spin up back. 
so that I can’t meet the SLA of 1 min. 

Kaniz
Community Manager
Community Manager

Hi @greyamber, Could you please configure it appropriately, so it will efficiently handle your REST API calls while meeting your SLA? Remember to adjust the cluster size based on your workload to strike the right balance between perfo...

Kaniz
Community Manager
Community Manager

Our End-of-Year Community Survey is here! Please take a few moments to complete the survey. Your feedback matters!

kulkpd
Contributor

@greyamber Interactive cluster costs two time more than job cluster. can you explain use-case of why job API needs to invoked and what API is doing. 

greyamber
New Contributor II

We need to pull data from an api endpoint at every 1 minute interval. Time to fetch and process is 30 seconds. 
if we use job cluster once the process completes it get terminated and for next run we have to wait ~3-5 mins to spin up the cluster. Which over pass our SLA 1 minute. 

Welcome to Databricks Community: Lets learn, network and celebrate together

Join our fast-growing data practitioner and expert community of 80K+ members, ready to discover, help and collaborate together while making meaningful connections. 

Click here to register and join today! 

Engage in exciting technical discussions, join a group with your peers and meet our Featured Members.