Bugzilla

(In reply to Marco Castelluccio [:marco] from comment #7)
> It's done on Taskcluster (not from mozilla-central CI though), the Taskcluster task is uploading to GCP. We could upload it as an artifact too, but downloading it from GCP might be better (so we avoid uploading the same file twice and pay for the storage twice).

Longer term, I think the ideal situation is that all of this (searchfox indexing, code coverage runs, searchfox analysis generation) happens inside of taskcluster and without involving external services.  In that case, I think the costs are acceptable.  (Uploading the artifact to S3 is free (in the same region), storage is $0.0023/GB/mo, and retrieval is free (same region).

We're not there yet though, so I think the big question is can you provide shell script logic that tells searchfox how to derive the URL for your GCP storage and fetch it inside the script at https://github.com/mozsearch/mozsearch-mozilla/blob/master/shared/fetch-tc-artifacts.sh.

(In reply to Marco Castelluccio [:marco] from comment #7)
> It's done on Taskcluster (not from mozilla-central CI though), the Taskcluster task is uploading to GCP. We could upload it as an artifact too, but downloading it from GCP might be better (so we avoid uploading the same file twice and pay for the storage twice).

Longer term, I think the ideal situation is that all of this (searchfox indexing, code coverage runs, searchfox analysis generation) happens inside of taskcluster and without involving external services.  In that case, I think the costs are acceptable.  (Uploading the artifact to S3 is free (in the same region), storage is $0.0023/GB/mo, and retrieval is free (same region).)

We're not there yet though, so I think the big question is can you provide shell script logic that tells searchfox how to derive the URL for your GCP storage and fetch it inside the script at https://github.com/mozsearch/mozsearch-mozilla/blob/master/shared/fetch-tc-artifacts.sh.