Currently, when using the Jib client to push images to Harbor registry, we are encountering a 502 error results in a complete failure
due to loss of http connection, requiring the entire push process to be manually restarted from scratch. This can be extremely disruptive, especially with huge images.
Expected behavior:
Successful push/pull operations.
Log output
Proposal
To address this issue, we propose implementing a retry mechanism (like docker and nerdctl) for Jib client push failures. This mechanism would automatically attempt to retry the push operation a specified number of times upon encountering a 502 error, significantly improving the overall reliability and robustness of the image pushing process.
Description of the issue:
Currently, when using the Jib client to push images to Harbor registry, we are encountering a 502 error results in a complete failure due to loss of http connection, requiring the entire push process to be manually restarted from scratch. This can be extremely disruptive, especially with huge images.
Expected behavior:
Successful push/pull operations.
Log output
Proposal
To address this issue, we propose implementing a retry mechanism (like docker and nerdctl) for Jib client push failures. This mechanism would automatically attempt to retry the push operation a specified number of times upon encountering a 502 error, significantly improving the overall reliability and robustness of the image pushing process.