Esctl

A Command-Line Interface designed to ease Elasticsearch administration.

Key Features • Installation • How To Use • Examples • License • Developing

---

Esctl is a CLI tool for Elasticsearch. I designed it to shorten huge curl commands Elasticsearch operators were running like :

curl -XPUT --user "john:doe" 'http://elasticsearch.example.com:9200/_cluster/settings' -d '{
    "transient" : {
        "cluster.routing.allocation.enable": "NONE"
    }
}'

The equivalent with esctl is

esctl cluster routing allocation enable none

Key Features

• Easy to use CLI rather than long curl commands (thanks to cliff)
• Cluster-level informations : stats, info, health, allocation explanation
• Node-level informations : list, hot threads, exclusion, stats
• Cluster-level and index-level settings
• _cat API for allocation, plugins and thread pools
• Index management : open, close, create, delete, list
• raw command to perform raw HTTP calls when esctl doesn't provide a nice interface for a given route.
• Per-module log configuration
• X-Pack APIs : users and roles
• Multiple output formats : table, csv, json, value, yaml
• JMESPath queries using the --jmespath flag
• Colored output !
• Run arbitrary pre-commands before issuing the call to Elasticsearch (like running kubectl port-forward for example)
• Fetch cluster's credentials from external commands instead of having them shown in cleartext in the config file

Installation

Using PIP

pip install esctl

From source

pip install git+https://github.com/jeromepin/esctl.git

How To Use

Esctl relies on a ~/.esctlrc file containing its config. This file is automatically created on the first start if it doesn't exists :

clusters:
  bar:
    servers:
      - https://bar.example.com

users:
  john-doe:
    username: john
    external_password:
      command:
        run: kubectl --context=bar --namespace=baz get secrets -o json my-secret | jq -r '.data.password||@base64d'

contexts:
  foo:
    user: john-doe
    cluster: bar

default-context: foo

Running pre-commands

Sometimes, you need to execute a shell command right before running the esctl command. Like running a kubectl port-forward in order to connect to your Kubernetes cluster. There is a pre_commands block inside the context which can take care of that :

clusters:
  remote-kubernetes:
    servers:
    - http://localhost:9200
contexts:
  my-distant-cluster:
    cluster: remote-kubernetes
    pre_commands:
    - command: kubectl --context=my-kubernetes-context --namespace=elasticsearch port-forward svc/elasticsearch 9200
      wait_for_exit: false
      wait_for_output: Forwarding from
    user: john-doe

Along with command, you can pass two options :

• wait_for_exit (default: true) : wait for the command to exit before continuing. Usually set to false when the command is running in the foreground.
• wait_for_output : if wait_for_exit is false, look for a specific output in the command's stdout. The string to look-for is interpreted as a regular expression passed to Python's re.compile().

Examples

License

esctl is licensed under the GNU GPLv3. See LICENCE file.

Developing

Install

make install

Run tests

make test

Format and lint code

make lint