本文已经收录在 Prometheus 合集 Prometheus 都可以采集那些指标？-- 常用 Exporter 合集 中。

Elasticsearch 是一个实时的分布式存储、搜索、分析的引擎，官网是 https://www.elastic.co/cn/elasticsearch/ 。 很多公司在用这个，你的公司大概率也有，那么对于 Elasticsearch 该怎么监控呢？针对这个问题，Prometheus Community 开发了一个 Elasticsearch Exporter ，用来导出 Elasticsearch 的监控数据，仓库地址是https://github.com/prometheus-community/elasticsearch_exporter ，用 Golang 编写。

Elasticsearch Exporter 当前版本是 1.3.0 ，发布于 2021.10.21

安装和配置

安装是比较简单的，下载最新版本的二进制包，或者下载 Docker image 。启动 Docker image 可以使用

docker run --rm -p 9114:9114 quay.io/prometheuscommunity/elasticsearch-exporter:latest --es.uri=http://localhost:9200 

在 Kubernetes 里使用 helm 部署可以参考 https://github.com/kubernetes/charts/tree/master/stable/elasticsearch-exporter 。

每次从这个 Exporter 获取 Elasticsearch 的监控数据都会从 Elasticsearch 服务拉取一次，如果获取时间间隔太短的话，可能会增大服务端的压力，尤其是使用 --es.all 和 --es.indices 运行 Exporter 的时候。所以在实际使用过程中，建议大家在确定获取数据间隔的时候，先测量获取 /_nodes/stats 和 /_all/_stats 指标的时间，防止监控对服务造成影响。

Elasticsearch Exporter 启动的时候有如下启动参数可以选择。

• es.uri Address (host and port) of the Elasticsearch node we should connect to. This could be a local node (localhost:9200, for instance), or the address of a remote Elasticsearch server. When basic auth is needed, specify as: <proto>://<user>:<password>@<host>:<port>. E.G., http://admin:pass@localhost:9200. Special characters in the user credentials need to be URL-encoded.

• es.all If true, query stats for all nodes in the cluster, rather than just the node we connect to.

• es.cluster_settings If true, query stats for cluster settings.

• es.indices If true, query stats for all indices in the cluster.

• es.indices_settings If true, query settings stats for all indices in the cluster.

• es.indices_mappings If true, query stats for mappings of all indices of the cluster. |false

• es.shards If true, query stats for all indices in the cluster, including shard-level stats (implies es.indices=true).

• es.snapshots If true, query stats for the cluster snapshots.

• es.timeout Timeout for trying to get stats from Elasticsearch. (ex: 20s)

• es.ca Path to PEM file that contains trusted Certificate Authorities for the Elasticsearch connection.

• es.client-private-key Path to PEM file that contains the private key for client auth when connecting to Elasticsearch.

• es.client-cert Path to PEM file that contains the corresponding cert for the private key to connect to Elasticsearch.

• es.clusterinfo.interval Cluster info update interval for the cluster label

• es.ssl-skip-verify Skip SSL verification when connecting to Elasticsearch.

• es.apiKey API Key to use for authenticating against Elasticsearch.

• web.listen-address Address to listen on for web interface and telemetry. 缺省是:9114
  

• web.telemetry-path Path under which to expose metrics. 缺省是 /metrics
  

• version Show version info on stdout and exit.

命令行参数以单个参数开始，用于版本低于 1.1.0rc1 。对于大于 1.1.0rc1 的版本，命令行参数用 -- 指定。此外，所有命令行参数都可以作为环境变量提供。

对于 Elasticsearch 7.x 版本提供了一些安全性的措施，支持 RBAC，用户名和密码可以直接在 URI 中传递，也可以通过`ES_USERNAME`和`ES_PASSWORD`环境变量传递。如果在 URI 中传递了身份验证，那么指定这两个环境变量以后将覆盖 URI 里的内容。

指标

Elasticsearch Exporter 可以导出很多指标，部分指标如下所示：

• elasticsearch_breakers_estimated_size_bytes | gauge | Estimated size in bytes of breaker

• elasticsearch_breakers_limit_size_bytes | gauge | Limit size in bytes for breaker

• elasticsearch_breakers_tripped | counter | tripped for breaker

• elasticsearch_cluster_health_active_primary_shards | gauge | The number of primary shards in your cluster. This is an aggregate total across all indices.

• elasticsearch_cluster_health_active_shards | gauge | Aggregate total of all shards across all indices, which includes replica shards.

• elasticsearch_cluster_health_delayed_unassigned_shards | gauge | Shards delayed to reduce reallocation overhead

• elasticsearch_cluster_health_initializing_shards | gauge | Count of shards that are being freshly created.

• elasticsearch_cluster_health_number_of_data_nodes | gauge | Number of data nodes in the cluster.

全部的指标大约有 120+ 条以上，详细指标可以参考 https://github.com/prometheus-community/elasticsearch_exporter#metrics 。