诺曼底:Firefox食谱服务器

  • h9_459476
    了解作者
  • 544.8KB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-04-25 01:05
    上传日期
诺曼底食谱服务器 诺曼底管理着对Firefox进行更改的方法,包括临时研究,用户调查和新功能的受控发布。 执照 诺曼底已获得MPLv2许可。 有关详细信息,请参见LICENSE文件。 为诺曼底发布的Docker映像包括由MaxMind创建的GeoLite2数据,可从。 骇客 请参阅文档以开始使用。
normandy-master.zip
内容介绍
Overview --- These files are appended to the hindsight configurations that produce parquet files in s3. Structure --- These files are organized by log stream. Log streams are named `${app}.${type}.${source}`, where `app` is always `normandy`, `type` is `app` or `admin`, and `source` is `docker.app`, but may change based on logging configuration. For each log stream there can be a `${log_stream}.cfg` file, or a directory `${log_stream}/` which contains .cfg files. All cfg files must contain a `parquet_schema`, and are interpreted as lua. If no cfg file is specified for a stream, then a fallback schema is used. Providing a file for a stream indicates that there is only one schema for the stream, and generally only requires specifying `parquet_schema`. This field accesses the message after PII is scrubbed. A string to string map of mozlog `Fields` is provided, and all values are also copied outside the map for use as schema columns, with field names modified by converting to lowercase, replacing `.`s with `_`, and adding a prefix of `fields_`. Providing a directory for a stream indicates that the stream is going to be split into multiple parquet schemas, so `message_matcher` and `s3_path_dimensions` should be specified. These fields access the message before PII is scrubbed, so metadata is referenced as `Hostname`, mozlog metadata is referenced as `Fields[Hostname]`, and mozlog fields are accessed like `Fields[Fields.agent]`. The default value of `message_matcher` will match the log stream for the file, so extending the matcher is preferrable using the lua `..` operator for string concatenation. In order to keep logs that don't match specific schemas, a fallback cfg should be provided that negates all the other message matchers in the directory, and uses the default mozlog parquet schema. Date and Hour are special fields that are extracted from the mozlog `Timestamp` value, for use in `s3_path_dimesions`. `s3_path_dimensions` is a list of partition names mapped to a hindsight `read_message()` source. It is standard to add a partition between `log` and `date` called `type` that matches the name of the cfg file. The default value partitions logs by log stream (heka message `Type`), `Date`, and `Hour` with: s3_path_dimensions = { {name="log", source="Type"}, {name="date", source="Fields[Date]"}, {name="hour", source="Fields[Hour]"}, }
评论
    相关推荐