WIP: Approach 1: a new lambda to demo a backoff-and-retry approach to SQS-driven processing

[phad]

Jira Ticket

DCMAW-XXXXX

Description of changes

Review guidance

Review checklist

Functional Review

• Functionality: Does it meet the acceptance criteria on the ticket and work as expected?
• Requirements: Does the code meet functional and non-functional requirements including compliance with programme standards and security gates?

Security & Compliance

• Personally Identifiable Information: Is it possible for PII to be logged?
• Security Considerations: Are there any security implications that need to be addressed?

Quality Assurance

• Testing: Is the code well-tested with sufficient coverage to provide confidence in correctness?
• Edge Cases: Have edge cases been considered and handled appropriately?

Code Quality

• Readability: Is the code easy to understand for all team members, with clear naming and appropriate documentation?
• Maintainability: Is the code easy to change, reuse, and extend?
• Code Style: Does it follow our coding conventions and best practices?
• Code Quality: Is the code maintainable and following best practices? See Values, Principles & Practices

Observability & Operations

• Observability: Are there appropriate logs/metrics that would help debug and monitor the service?
• Performance: Are there any performance considerations or potential bottlenecks?
• Runbooks for Alarms: If an alarm has been created or updated, has a corresponding runbook been created or updated?
  • Critical alarms runbook
  • Warning alarms runbook

Documentation

• Documentation: Is the code well documented? Is there any existing documentation that needs updating?
• Comments: Are complex sections of code adequately commented if the intent is not clear?

Review PR:

• Title: Contains ticket number and clear summary of change
• Description: Has clear description of change

Evidence

Documentation

[sandymay-dd]

initialDelaySec becomes a bit misleading when it's being used in the backoff calculations for subsequent retries.

[phad]

Yes, maybe - this is the 'absolute' calculation mentioned in https://govukverify.atlassian.net/wiki/spaces/DCMAW/pages/5950603379/2015-12-08+Investigate+Define+ReadID+Polling+Strategy#Absolute so it is technically correct that it's the initial back-off period, multiplied by the back-off factor ^ try index.