Decide the approach for post-deployment smoke testing on UAT/PROD
- Dominant language
- Python
- Stars
- 4
- Forks
- 0
- Avg merge
- 1d 5h
- Merged PRs (30d)
- 8
Description
### Background
database/migrations/V1.4.0.2__initial_schema.ddl includes a public_cps_za_test table used to write/read a "test topic" event, originally added by @oto-macenauer-absa as a PoC mechanism for basic post-deployment smoke testing (write to a test topic, read via SQS, confirm the pipeline works end-to-end). New question was raised whether this is still needed now that a DEV environment exists for testing, and flagged that it means a test-only object lives permanently inside the real UAT/PROD schema. @miroslavpojer raised that running any automated smoke test — read or write — against PROD carries real risk, especially write-based tests, which could be silently forgotten and cause harm over time.
### Questions To Answer
1. Do we still need an automated post-deployment smoke test against UAT/PROD, given DEV exists for general testing?
2. If yes, should it be read-only, write-based, or both?
3. Should smoke-test objects (tables/topics) be kept separate from the real production schema/queues instead of embedded alongside real data (current solution)?
4. Should public_cps_za_test be removed, kept as-is, or replaced with a differently scoped mechanism?
### Additional Resources
• Discussion thread https://github.com/AbsaOSS/EventGate/pull/214#discussion_r3813778453.
• Related: #201 (Flyway migrations setup)
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.