Makes sense — idempotence testing is a great use case.
We've hit bugs where retries created duplicate charges.
If this existed as a hosted tool with a UI to configure
scenarios (instead of encoding in txids), would that
have saved you time? Or was code-based actually preferred?
proxy that does all of the above rather than mocks. we had configurable one based on txid.
ex: tx uuid fe83dd01-aaaa-aaaa-aaaa-000000005373 if chunk1 is aaaa, then treat last bit as 'delay for 5373ms'
Clever approach — encoding the behavior in the txid itself.
Did you have a UI to configure scenarios, or was it all code-based? And did you simulate webhook failures/retries too, or just delays?
it was code base. first chunk would be git revision-sha . failure, retry, delay, multiple replay due to retry timeout (idempotence test)
Makes sense — idempotence testing is a great use case. We've hit bugs where retries created duplicate charges.
If this existed as a hosted tool with a UI to configure scenarios (instead of encoding in txids), would that have saved you time? Or was code-based actually preferred?
dont care would never have used hosted.