We are currently getting some of our build logs truncated to 5MB and was wondering if setting DRONE_S3_BUCKET would help eliminate this limitation? I poked around internally and another team is using S3 and the largest file I see in there is about 8MB, but it’s all json-encoded logs and the first log shows a time field of 176, which tells me that it too is getting truncated. Using jq to just output the raw logs, that comes out to about 5MB of total size. How can we increase the allowed size of our build logs?
1 Like
FWIW, I wrote my large logs to a file and uploaded it elsewhere to see the raw output and it was just over 15MB. It’s output from a large terragrunt run across a huge number of terraform repos … not exactly something we can split up into multiple steps and not something we can tolerate truncated logs for.
I just hit this too. We have a pipeline for integration tests, with a couple of web servers running as services in that pipeline. The tests are failing due to errors in those servers. These occur early on in the pipeline, and so we have no way of finding out what the errors are because they are not in the last 5MB of the output for those services.