As part of our upgrade from ats6.2 to ats9, we ran into a problem with elevated ERR_CLIENT_ABORT counts. After a lot of investigation, it turned out to be not a regression or due to a functional problem, but just due to a cause code change with PR #4267 (#4267). The same rate of errors have been there before too but were being put into other buckets such as ERR_UNKNOWN and are just shuffled around. I confirmed this by removing this specific patch that the ERR_CLIENT_ABORTS went back to our older baseline levels.
Creating this issue so others that may run into a similar problem will not have to spend time troubleshooting this.
As part of our upgrade from ats6.2 to ats9, we ran into a problem with elevated ERR_CLIENT_ABORT counts. After a lot of investigation, it turned out to be not a regression or due to a functional problem, but just due to a cause code change with PR #4267 (#4267). The same rate of errors have been there before too but were being put into other buckets such as ERR_UNKNOWN and are just shuffled around. I confirmed this by removing this specific patch that the ERR_CLIENT_ABORTS went back to our older baseline levels.
Creating this issue so others that may run into a similar problem will not have to spend time troubleshooting this.