An ops discovery of serverless
Download as PPTX, PDF1 like72 views
The document discusses using serverless computing to schedule long-running or compute-intensive tasks that need to execute only once across a cluster. It presents two scenarios: 1) using CloudWatch events to trigger a Lambda function that puts a task into an SQS queue when it needs to be run, and 2) using CloudWatch metrics to monitor the status of pending tasks by querying a database from a Lambda function. The document is a presentation on exploring serverless options for operations tasks.
An ops discovery of serverless
- 1. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. BENGALURU
- 2. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. An Ops discovery of Serverless Ninad Pundalik | 6 October 2018
- 3. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. $ which
- 4. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Data store + Application Scaling Group
- 5. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Scenario I What were we trying to do? ● Schedule a task ● which definitely needs to execute only once across the entire cluster ● that may be long running? ● probably compute intensive too...
- 6. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. CW Event Lambda SQS
- 7. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Scenario II Monitor the status of pending tasks by querying a DB table (or whatever your data store calls a collection)
- 8. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. CW Event Lambda CW Metrics
- 9. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Questions?
- 10. © 2018, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Thank you! Ninad Pundalik | Astro Technology Inc (acquired by Slack) | 6 October 2018