Sentry 查询基本逻辑

查询api的基本查询调用逻辑如下：

查询api()

=> prep_search() [执行查询]

=> build_query_params_from_request() : { projects, sort_by: sort, limit, cursor, search_filters: query} [构建查询参数]

=> parse_search_query() : search_filters [解析查询字符串]

=> search.query() ：result [查询结果]

=> snuba_sdk.query.Query() : [执行具体查询]

search_filters是一个由 query 参数格式化来的查询器集合。一切未在接口定义中明确定义且符合sentry 查询格式的参数均可以通过这个查询器的校验，并执行查询。

根据时间范围查询

parse_search_query方法根据语法解析查询字符串，判断查询条件是否正确，关于时间的查询器的定义如下：

# filter for dates
date_filter = search_key sep operator iso_8601_date_format

key                  = ~r"[a-zA-Z0-9_.-]+"
quoted_key           = '"' ~r"[a-zA-Z0-9_.:-]+" '"'
search_key           = key / quoted_key
iso_8601_date_format = date_format time_format? ("Z" / tz_format)? &end_value
sep                  = ":"
operator             = ">=" / "<=" / ">" / "<" / "=" / "!="

search.query方法会将search_filters中的date和另两个参数date_from、date_to比较，得到最终的查询时间范围。

1
2
3

end_params = [_f for _f in [date_to, get_search_filter(search_filters, "date", "<")] if _f]
# other code
start_params = [date_from, retention_date, get_search_filter(search_filters, "date", ">")]

其中 get_search_filter 方法根据查询参数名和操作符获取查询参数值，如果找到多个值，则返回最严格的值。定义如下

def get_search_filter(search_filters: Sequence[SearchFilter], name: str, operator: str) -> Any:
    """
    Finds the value of a search filter with the passed name and operator. If
    multiple values are found, returns the most restrictive value
    :param search_filters: collection of `SearchFilter` objects
    :param name: Name of the field to find
    :param operator: '<', '>' or '='
    :return: The value of the field if found, else None
    """
    if not search_filters:
        return None
    assert operator in ("<", ">", "=", "IN")
    comparator = max if operator.startswith(">") else min
    found_val = None
    for search_filter in search_filters:
        # Note that we check operator with `startswith` here so that we handle
        # <, <=, >, >=
        if search_filter.key.name == name and search_filter.operator.startswith(operator):
            val = search_filter.value.raw_value
            found_val = comparator(val, found_val) if found_val else val
    return found_val