fix(amazonq): retry S3 upload #5208
Conversation
| } | ||
| LOG.info { "Upload to S3 succeeded" } |
There was a problem hiding this comment.
I'm not a big fan of implementing our own retry logic rather than doing it at the SDK layer, but the existing S3 client's upload functionality requires a bucket name as a part of the request, and 1) I don't think we should have the internal S3 bucket names we use floating around this repo, 2) we use multiple different buckets, and 3) we don't have access to the bucket name in the response of our CreateUploadUrl — this API would have to be changed and then we can use the existing S3 client.
There was a problem hiding this comment.
Update: use waitUntil instead of a for loop; makes this retry logic even simpler
| UnknownHostException::class, | ||
| SocketTimeoutException::class, | ||
| HttpRequests.HttpStatusException::class, | ||
| ConnectException::class |
There was a problem hiding this comment.
UnknownHostException and SocketTimeoutException can be reproduced by turning off your WiFi, HttpStatusException could be a 500 or something, and I've occasionally seen ConnectException too, so I think all of these are worth retrying on.
Types of changes
Description
Retry S3 upload.
Checklist
License
I confirm that my contribution is made under the terms of the Apache 2.0 license.