# profile接口

profiler 是 Elasticsearch 5.0 的一个新接口。通过这个功能，可以看到一个搜索聚合请求，是如何拆分成底层的 Lucene 请求，并且显示每部分的耗时情况。

启用 profiler 的方式很简单，直接在请求里加一行即可：

```
curl -XPOST 'http://localhost:9200/_search' -d '{
    "profile": true,
    "query": { ... },
    "aggs": { ... }
}'
```

可以看到其中对 query 和 aggs 部分的返回是不太一样的。

## query

query 部分包括 collectors、rewrite 和 query 部分。对复杂 query，profiler 会拆分 query 成多个基础的 TermQuery，然后每个 TermQuery 再显示各自的分阶段耗时如下：

```
"breakdown": {
    "score": 51306,
    "score_count": 4,
    "build_scorer": 2935582,
    "build_scorer_count": 1,
    "match": 0,
    "match_count": 0,
    "create_weight": 919297,
    "create_weight_count": 1,
    "next_doc": 53876,
    "next_doc_count": 5,
    "advance": 0,
    "advance_count": 0
}
```

## aggs

```
        "time": "1124.864392ms",
        "breakdown": {
            "reduce": 0,
            "reduce_count": 0,
            "build_aggregation": 1394,
            "build_aggregation_count": 150,
            "initialise": 2883,
            "initialize_count": 150,
            "collect": 1124860115,
            "collect_count": 900
        }
```

我们可以很明显的看到聚合统计在初始化阶段、收集阶段、构建阶段、汇总阶段分别花了多少时间，遍历了多少数据。

*注意其中 reduce 阶段还没实现完毕，所有都是 0。因为目前 profiler 只能在 shard 级别上做统计。*

collect 阶段的耗时，有助于我们调整对应 aggs 的 `collect_mode` 参数选择。目前 Elasticsearch 支持 `breadth_first` 和 `depth_first` 两种方式。

initialise 阶段的耗时，有助于我们调整对应 aggs 的 `execution_hint` 参数选择。目前 Elasticsearch 支持 `map`、`global_ordinals_low_cardinality`、`global_ordinals` 和 `global_ordinals_hash` 四种选择。在计算离散度比较大的字段统计值时，适当调整该参数，有益于节省内存和提高计算速度。

*对高离散度字段值统计性能很关注的读者，可以关注* <https://github.com/elastic/elasticsearch/pull/21626> *这条记录的进展。*


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://hezhiqiang.gitbook.io/elkstack/elasticsearch/xing-neng-you-hua/profile.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.