Why API key management in Elasticsearch? - Purpose & Use Cases

The Big Idea

What if one lost password could put your entire Elasticsearch data at risk?

The Scenario

Imagine you have to give many people access to your Elasticsearch data. You write down their usernames and passwords on paper or in a simple file. Every time someone needs access, you share the password manually.

The Problem

This manual way is slow and risky. Passwords can be lost, shared with the wrong people, or forgotten. Changing access means updating everyone one by one. It's easy to make mistakes and hard to keep your data safe.

The Solution

API key management lets you create special keys for each user or app. You can give limited access, set expiration times, and revoke keys anytime. This keeps your data safe and makes access easy to control.

Before vs After

✗ Before

curl -u user:password 'http://localhost:9200/_search'
curl -u user:password 'http://localhost:9200/_search'
# Password shared everywhere

✓ After

curl -H 'Authorization: ApiKey <encoded_key>' 'http://localhost:9200/_search'
curl -H 'Authorization: ApiKey <encoded_key>' 'http://localhost:9200/_search'
# Unique keys per user, easy to revoke

What It Enables

You can safely share access to Elasticsearch with many users or apps, controlling who can do what and when.

Real Life Example

A company gives each developer a unique API key to access Elasticsearch logs. If a key is lost or a developer leaves, the key can be revoked without changing everyone else's access.

Key Takeaways

Manual password sharing is risky and hard to manage.

API key management creates secure, controllable access keys.

It makes sharing and revoking access simple and safe.

Practice

1. What is the primary purpose of an API key in Elasticsearch?

easy

A. To monitor Elasticsearch cluster health

B. To store data inside Elasticsearch indices

C. To allow applications to securely access Elasticsearch with specific permissions

D. To backup Elasticsearch data automatically

Solution

1. Step 1: Understand API key role

   API keys are secret tokens used to authenticate and authorize applications.

2. Step 2: Identify purpose in Elasticsearch

   They grant controlled access to Elasticsearch resources based on assigned roles.

3. Final Answer:

   To allow applications to securely access Elasticsearch with specific permissions -> Option C

4. Quick Check:

   API key = secure app access [OK]

Hint: API keys control app access permissions [OK]

Common Mistakes:

• Confusing API keys with data storage
• Thinking API keys monitor cluster health
• Assuming API keys handle backups

2. Which of the following is the correct Elasticsearch API call to create an API key?

easy

A. DELETE /_security/api_key

B. GET /_security/api_key/create

C. PUT /_security/api_key

D. POST /_security/api_key

Solution

1. Step 1: Recall API key creation syntax

   Elasticsearch uses POST method to create resources like API keys.

2. Step 2: Match correct endpoint

   The correct endpoint for creating an API key is POST /_security/api_key.

3. Final Answer:

   POST /_security/api_key -> Option D

4. Quick Check:

   POST + /_security/api_key = create key [OK]

Hint: Use POST to create API keys in Elasticsearch [OK]

Common Mistakes:

• Using GET or DELETE for creation
• Confusing endpoint paths
• Using PUT instead of POST

3. Given this API key creation request body, what will be the name of the created API key?

{
  "name": "my-app-key",
  "role_descriptors": {
    "my-role": {
      "cluster": ["all"],
      "index": [{"names": ["logs-*"], "privileges": ["read"]}]
    }
  }
}

medium

A. my-app-key

B. my-role

C. logs-*

D. all

Solution

1. Step 1: Identify the API key name field

   The "name" field in the request body sets the API key's name.

2. Step 2: Read the value of the "name" field

   The value is "my-app-key", which becomes the API key's name.

3. Final Answer:

   my-app-key -> Option A

4. Quick Check:

   API key name = "name" field value [OK]

Hint: API key name is in the "name" field [OK]

Common Mistakes:

• Confusing role name with API key name
• Using index pattern as key name
• Mistaking privileges for name

4. You try to delete an API key using this request: DELETE /_security/api_key?id=12345 but get an error. What is the likely cause?

medium

A. API key names cannot be deleted, only IDs

B. API key ID must be passed in the request body, not as a query parameter

C. DELETE method is not supported for API keys

D. You must use GET method to delete API keys

Solution

1. Step 1: Check API key deletion syntax

   Elasticsearch requires the API key ID in the request body JSON, not as a URL query parameter.

2. Step 2: Understand method support

   DELETE method is supported, but parameters must be correctly passed in the body.

3. Final Answer:

   API key ID must be passed in the request body, not as a query parameter -> Option B

4. Quick Check:

   Delete API key ID in body, not URL [OK]

Hint: Pass API key ID in JSON body for deletion [OK]

Common Mistakes:

• Passing ID as URL query parameter
• Using wrong HTTP method
• Confusing API key name with ID

5. You want to create an API key that only allows reading from indices starting with "sales-" and no cluster privileges. Which role descriptor is correct in the request body?

hard

A. { "role_descriptors": { "read_sales": { "cluster": [], "index": [{ "names": ["sales-*"], "privileges": ["read"] }] } } }

B. { "role_descriptors": { "read_sales": { "cluster": ["all"], "index": [{ "names": ["sales-*"], "privileges": ["write"] }] } } }

C. { "role_descriptors": { "read_sales": { "cluster": ["monitor"], "index": [{ "names": ["sales-*"], "privileges": ["all"] }] } } }

D. { "role_descriptors": { "read_sales": { "cluster": ["all"], "index": [{ "names": ["*"], "privileges": ["read"] }] } } }

Solution

1. Step 1: Identify required privileges

   The API key should have no cluster privileges and only read privileges on indices starting with "sales-".

2. Step 2: Match role descriptor to requirements

   { "role_descriptors": { "read_sales": { "cluster": [], "index": [{ "names": ["sales-*"], "privileges": ["read"] }] } } } has empty cluster privileges and read privilege on "sales-*" indices, matching the requirement.

3. Final Answer:

   { "role_descriptors": { "read_sales": { "cluster": [], "index": [{ "names": ["sales-*"], "privileges": ["read"] }] } } } -> Option A

4. Quick Check:

   No cluster + read sales-* = { "role_descriptors": { "read_sales": { "cluster": [], "index": [{ "names": ["sales-*"], "privileges": ["read"] }] } } } [OK]

Hint: Empty cluster array means no cluster privileges [OK]

Common Mistakes:

• Giving cluster all privileges by mistake
• Using write or all privileges instead of read
• Applying privileges to wrong index patterns