Reduce the size of the blobs index
Per https://gitlab.com/gitlab-com/infrastructure/issues/1597#note_39253531
On GitLab.com, the elasticsearch index is predicted to be absolutely huge. Initially I thought this was dominated by commits, but that doesn't seem to be case.
We have https://gitlab.com/gitlab-org/gitlab-ee/issues/3217 and https://gitlab.com/gitlab-org/gitlab-ee/issues/2225 to consider ways to reduce the size of the commit documents. Some of the suggestions could apply equally to blobs.
Is there anything else we could do for blobs specifically? Are our analyzers creating grossly bloated indexes? Anything else?
/cc @smcgivern @vsizov
Edited by Nick Thomas