Role-based access control in Elasticsearch - Time & Space Complexity

Time Complexity: Role-based access control

O(n)

Understanding Time Complexity

When using role-based access control in Elasticsearch, it is important to understand how the system checks permissions as the number of roles and users grows.

We want to know how the time to verify access changes when more roles or permissions are added.

Scenario Under Consideration

Analyze the time complexity of the following role check query.

POST /_security/user/_has_privileges
{
  "username": "alice",
  "privileges": {
    "cluster": ["monitor"],
    "index": [
      {
        "names": ["sales-data"],
        "privileges": ["read"]
      }
    ]
  }
}
    

This request checks if user "alice" has the specified cluster and index privileges based on her assigned roles.

Identify Repeating Operations

When Elasticsearch checks privileges, it:

• Primary operation: Iterates over all roles assigned to the user.
• How many times: Once per role, checking each role's permissions.

The dominant work is checking each role's permissions against requested privileges.

How Execution Grows With Input

As the number of roles assigned to a user increases, the time to check privileges grows roughly in proportion.

Input Size (number of roles)	Approx. Operations
10	10 permission checks
100	100 permission checks
1000	1000 permission checks

Pattern observation: The time grows linearly as more roles are checked.

Final Time Complexity

Time Complexity: O(n)

This means the time to verify access grows linearly with the number of roles assigned to the user.

Common Mistake

[X] Wrong: "Checking user privileges is always constant time regardless of roles."

[OK] Correct: The system must check each role's permissions, so more roles mean more checks and longer time.

Interview Connect

Understanding how role checks scale helps you explain system performance and design better access control in real projects.

Self-Check

What if we changed from checking roles one by one to caching combined permissions? How would the time complexity change?

Practice

1. What is the main purpose of Role-based Access Control (RBAC) in Elasticsearch?

easy

A. To control who can perform specific actions by assigning roles

B. To speed up search queries

C. To store data in different formats

D. To backup Elasticsearch clusters automatically

Solution

1. Step 1: Understand RBAC concept

   RBAC is about managing permissions by assigning roles to users.

2. Step 2: Identify RBAC purpose in Elasticsearch

   It controls who can do what actions on the cluster or indexes.

3. Final Answer:

   To control who can perform specific actions by assigning roles -> Option A

4. Quick Check:

   RBAC = Control access by roles [OK]

Hint: RBAC means controlling access by roles, not data or speed [OK]

Common Mistakes:

• Confusing RBAC with data storage or backup
• Thinking RBAC speeds up queries
• Assuming RBAC changes data formats

2. Which of the following is the correct JSON structure to define a role with read access to the index logs-2024?

easy

A. {"cluster": ["all"], "indices": [{"names": ["logs-2024"], "privileges": ["monitor"]}]}

B. {"cluster": ["all"], "indices": [{"names": ["logs-2024"], "privileges": ["write"]}]}

C. {"cluster": ["read"], "indices": [{"names": ["logs-2024"], "privileges": ["write"]}]}

D. {"cluster": ["monitor"], "indices": [{"names": ["logs-2024"], "privileges": ["read"]}]}

Solution

1. Step 1: Check cluster privileges for read access

   Read access to an index usually requires cluster privileges like 'monitor', not 'all' or 'read'.

2. Step 2: Verify index privileges

   The index privileges must include 'read' for the specified index.

3. Final Answer:

   {"cluster": ["monitor"], "indices": [{"names": ["logs-2024"], "privileges": ["read"]}]} -> Option D

4. Quick Check:

   Cluster 'monitor' + index 'read' = correct role [OK]

Hint: Cluster 'monitor' + index 'read' grants read access [OK]

Common Mistakes:

• Using 'all' cluster privilege unnecessarily
• Confusing 'write' with 'read' privileges
• Assigning 'read' cluster privilege which is invalid

3. Given this role definition, what permissions does a user have on the sales-data index?

{
  "cluster": ["monitor"],
  "indices": [
    {
      "names": ["sales-data"],
      "privileges": ["read", "write"]
    }
  ]
}

medium

A. User can read and write data in sales-data index

B. User can only read data from sales-data index

C. User can manage cluster settings but not access sales-data

D. User has full admin access to all indexes

Solution

1. Step 1: Analyze cluster privileges

   Cluster privilege 'monitor' allows monitoring but no write or admin cluster changes.

2. Step 2: Analyze index privileges

   Privileges 'read' and 'write' on 'sales-data' index allow reading and writing data there.

3. Final Answer:

   User can read and write data in sales-data index -> Option A

4. Quick Check:

   Index 'read' + 'write' = read/write access [OK]

Hint: Check index privileges for read/write to know access level [OK]

Common Mistakes:

• Ignoring 'write' privilege and assuming read-only
• Confusing cluster 'monitor' with admin rights
• Assuming full admin access without 'all' privilege

4. You defined this role but users report they cannot write to the app-logs index. What is the error?

{
  "cluster": ["monitor"],
  "indices": [
    {
      "names": ["app-logs"],
      "privileges": ["read"]
    }
  ]
}

medium

A. The cluster privilege 'monitor' is incorrect for write access

B. The index privilege should include 'write' to allow writing

C. The index name 'app-logs' is misspelled

D. The role JSON is missing a 'run_as' field

Solution

1. Step 1: Check index privileges

   The role only grants 'read' privilege on 'app-logs', so writing is not allowed.

2. Step 2: Identify missing privilege

   To write, the 'write' privilege must be added to the index privileges.

3. Final Answer:

   The index privilege should include 'write' to allow writing -> Option B

4. Quick Check:

   Write access needs 'write' privilege [OK]

Hint: Write access requires 'write' privilege on index [OK]

Common Mistakes:

• Assuming 'monitor' cluster privilege allows writing
• Overlooking missing 'write' privilege on index
• Thinking 'run_as' is required for write permission

5. You want to create a role that allows a user to read from all indexes starting with prod- but only write to prod-logs. Which role definition is correct?

hard

A. {
"cluster": ["all"],
"indices": [
{"names": ["prod-logs"], "privileges": ["read", "write"]}
]
}

B. {
"cluster": ["monitor"],
"indices": [
{"names": ["prod-logs"], "privileges": ["read", "write"]},
{"names": ["prod-*"], "privileges": ["read", "write"]}
]
}

C. {
"cluster": ["monitor"],
"indices": [
{"names": ["prod-*"], "privileges": ["read"]},
{"names": ["prod-logs"], "privileges": ["write"]}
]
}

D. {
"cluster": ["monitor"],
"indices": [
{"names": ["prod-*"], "privileges": ["write"]}
]
}

Solution

1. Step 1: Understand the requirement

   User needs read access on all 'prod-*' indexes and write only on 'prod-logs'.

2. Step 2: Check each option

   { "cluster": ["monitor"], "indices": [ {"names": ["prod-*"], "privileges": ["read"]}, {"names": ["prod-logs"], "privileges": ["write"]} ] } correctly assigns 'read' to 'prod-*' and 'write' to 'prod-logs'. { "cluster": ["all"], "indices": [ {"names": ["prod-logs"], "privileges": ["read", "write"]} ] } gives full cluster 'all' which is too broad. { "cluster": ["monitor"], "indices": [ {"names": ["prod-logs"], "privileges": ["read", "write"]}, {"names": ["prod-*"], "privileges": ["read", "write"]} ] } incorrectly grants 'read' and 'write' to all 'prod-*' indexes. { "cluster": ["monitor"], "indices": [ {"names": ["prod-*"], "privileges": ["write"]} ] } wrongly gives 'write' to all 'prod-*' indexes.

3. Final Answer:

   { "cluster": ["monitor"], "indices": [ {"names": ["prod-*"], "privileges": ["read"]}, {"names": ["prod-logs"], "privileges": ["write"]} ] } -> Option C

4. Quick Check:

   Read on prod-* + write on prod-logs = correct role [OK]

Hint: Use wildcard for read, specific index for write [OK]

Common Mistakes:

• Giving write privilege to all prod-* indexes
• Using cluster 'all' unnecessarily
• Mixing up index names and privileges