Shard sizing strategy in Elasticsearch - Time & Space Complexity

Time Complexity: Shard sizing strategy

O(n)

Understanding Time Complexity

When working with Elasticsearch, how we size shards affects how fast queries and indexing run.

We want to understand how the number and size of shards impact the work Elasticsearch does.

Scenario Under Consideration

Analyze the time complexity of querying data distributed across shards.

GET /my_index/_search
{
  "query": { "match_all": {} },
  "size": 10
}
    

This query searches all shards of the index and combines results.

Identify Repeating Operations

Look at what repeats when Elasticsearch runs this query.

• Primary operation: Searching each shard separately.
• How many times: Once per shard in the index.

How Execution Grows With Input

As you add more shards, Elasticsearch does more separate searches.

Number of Shards (n)	Approx. Searches
5	5 searches
50	50 searches
500	500 searches

Pattern observation: The work grows directly with the number of shards.

Final Time Complexity

Time Complexity: O(n)

This means the total work grows in a straight line as you add more shards.

Common Mistake

[X] Wrong: "More shards always make queries faster because work is split more."

[OK] Correct: Each shard adds overhead, so too many shards can slow things down instead of speeding them up.

Interview Connect

Understanding shard sizing helps you design Elasticsearch setups that balance speed and resource use, a key skill for real projects.

Self-Check

"What if we reduce the number of shards but increase the size of each shard? How would the time complexity change?"

Practice

1. What is the main reason to choose an appropriate shard size in Elasticsearch?

easy

A. To balance data storage and search performance

B. To increase the number of replicas

C. To reduce the number of indices

D. To avoid using any replicas

Solution

1. Step 1: Understand shard purpose

   Shards split data to distribute storage and speed up search operations.

2. Step 2: Connect shard size to performance

   Choosing the right shard size balances storage efficiency and search speed.

3. Final Answer:

   To balance data storage and search performance -> Option A

4. Quick Check:

   Shard size affects performance balance = A [OK]

Hint: Shard size balances storage and speed [OK]

Common Mistakes:

• Thinking replicas control shard size
• Confusing shard count with replica count
• Assuming more shards always improve speed

2. Which setting controls the number of primary shards when creating an Elasticsearch index?

easy

A. number_of_shards

B. number_of_replicas

C. shard_size

D. index_refresh_interval

Solution

1. Step 1: Identify shard count setting

   The setting number_of_shards defines how many primary shards an index has.

2. Step 2: Differentiate from replicas

   number_of_replicas controls copies, not primary shard count.

3. Final Answer:

   number_of_shards -> Option A

4. Quick Check:

   Primary shards = number_of_shards [OK]

Hint: Primary shards set by number_of_shards [OK]

Common Mistakes:

• Confusing replicas with shards
• Using shard_size which is not a setting
• Mixing index refresh with shard count

3. Given an index with 5 primary shards and each shard sized at 20GB, what is the total data size stored in the index?

medium

A. 20GB

B. 100GB

C. 25GB

D. 5GB

Solution

1. Step 1: Calculate total size from shards

   Total size = number of shards x size per shard = 5 x 20GB = 100GB.

2. Step 2: Confirm no replicas included

   Replicas add copies but do not affect primary data size calculation here.

3. Final Answer:

   100GB -> Option B

4. Quick Check:

   5 shards x 20GB = 100GB [OK]

Hint: Multiply shards by shard size [OK]

Common Mistakes:

• Adding replica size to primary data size
• Confusing shard count with replica count
• Choosing shard size instead of total

4. You set number_of_shards to 1 but your data size grows to 200GB. What is the main problem with this shard sizing?

medium

A. Index refresh interval is too short

B. Too many shards causing overhead

C. Replica count is zero

D. Shard size is too large, causing slower search and indexing

Solution

1. Step 1: Analyze shard size impact

   One shard holding 200GB is large and can slow down search and indexing.

2. Step 2: Identify correct problem

   Too few shards for large data causes performance issues, not replica count or refresh interval.

3. Final Answer:

   Shard size is too large, causing slower search and indexing -> Option D

4. Quick Check:

   Large shard size = slower performance [OK]

Hint: Avoid very large single shards [OK]

Common Mistakes:

• Blaming replica count instead of shard size
• Thinking many shards cause this problem
• Ignoring shard size impact on speed

5. You have 500GB of data and want to keep shard sizes between 10GB and 40GB. Which shard count is best to set for your index?

hard

A. 5 shards

B. 10 shards

C. 50 shards

D. 100 shards

Solution

1. Step 1: Calculate shard count range

   Minimum shards = 500GB / 40GB ≈ 13 shards; maximum shards = 500GB / 10GB = 50 shards.

2. Step 2: Choose shard count within range

   To keep shard size between 10GB and 40GB, choose a shard count near 50.

3. Final Answer:

   50 shards -> Option C

4. Quick Check:

   500GB ÷ 50 shards = 10GB per shard [OK]

Hint: Divide total data by desired shard size [OK]

Common Mistakes:

• Choosing too few shards causing large shard size
• Choosing too many shards causing overhead
• Ignoring shard size limits