redpanda-data / redpanda-data/connect
sftp output does not retry after a failed transfer operation. (Connection failed)
Nobody has claimed this yet.
- Dominant language
- Go
- Stars
- 8.8k
- Forks
- 969
- Avg merge
- 1d 13h
- Merged PRs (30d)
- 64
Description
Version 4.61.0:
It seems that when there is an issue with the sftp, say a permissions issue on overwriting a file, the sftp output will never reconnect to the sftp server. First there is a permission denied issue overwriting a file on the remote:
time="2025-08-12T17:45:56Z" level=error msg="Failed to send message to sftp: failed to open remote file: permission denied" @service=redpanda-connect label="" path=root.output.broker.outputs.0 stream=sftp
Then after the failure I start seeing these messages in the log:
time="2025-08-12T17:45:57Z" level=error msg="Failed to send message to sftp: failed to create SFTP client: ssh: rejected: connect failed (open failed)" @service=redpanda-connect label="" path=root.output.broker.outputs.0 stream=sftp
And on and on forever until this morning:
time="2025-08-13T16:21:14Z" level=error msg="Failed to send message to sftp: failed to create SFTP client: ssh: rejected: connect failed (open failed)" @service=redpanda-connect label="" path=root.output.broker.outputs.0 stream=sftp
If I restart my redpanda-connect pod, transfers resume normally. I believe this is a bug with the reuse of the ssh connection in the sftp output.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the SFTP output implementation and tracing how it reuses the SSH connection after a failed remote file operation. Reproduce the permission-denied transfer failure, then verify that the output reconnects and resumes transfers without restarting the redpanda-connect pod.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- backend, stream-processing
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100