14. Crash storage: AWS S3

The collector saves raw crash data to Amazon S3.

The processor loads raw crash data from Amazon S3, processes it, and then saves the processed crash data back to Amazon S3.

All of this is done in a single S3 bucket.

The “directory” hierarchy of that bucket looks like this:

  • {prefix}/v2/{name_of_thing}/{entropy}/{date}/{id}: Raw crash data.
  • {prefix}/v1/{name_of_thing}/{id}: Processed crash data, dumps, dump_names, and other things.