Cache management (query, request, field data) in Elasticsearch - Step-by-Step Execution

Concept Flow - Cache management (query, request, field data)

Receive Query Request

↓

Check Query Cache

↓

Return Cached

↓

Return Cached

↓

Return Cached

↓

Return Data

When a query request arrives, Elasticsearch checks caches in order: query cache, request cache, then field data cache. If a cache hit occurs, it returns cached data; otherwise, it loads data and caches it.

Execution Sample

Elasticsearch

GET /my_index/_search
{
  "query": { "match": { "field": "value" } }
}

This query searches 'my_index' for documents where 'field' matches 'value'. Elasticsearch checks caches before executing.

Execution Table

Step	Cache Checked	Cache Hit?	Action Taken	Result
1	Query Cache	No	Check next cache	Continue to Request Cache
2	Request Cache	No	Check next cache	Continue to Field Data Cache
3	Field Data Cache	No	Load data and cache	Data loaded and cached
4	Return Data	N/A	Return loaded data	Query result returned
5	Query Cache	Yes	Return cached query result	Cached query result returned

💡 Execution stops when cached data is found or data is loaded and returned.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	Final
query_cache	empty	miss	miss	miss	updated with query result
request_cache	empty	N/A	miss	miss	updated with request data
field_data_cache	empty	N/A	N/A	miss	updated with field data
response	none	none	none	loaded data	cached or loaded data

Key Moments - 3 Insights

Why does Elasticsearch check multiple caches instead of just one?

Because each cache stores different data types (query results, request results, field data), Elasticsearch checks them in order to find the fastest available data, as shown in steps 1-3 in the execution_table.

What happens if all caches miss?

Elasticsearch loads the data fresh from disk or memory, then updates the caches with this data, as shown in step 3 and 4 of the execution_table.

How does a cache hit improve performance?

A cache hit returns precomputed data immediately without re-executing the query or loading data again, as seen in step 5 where the query cache hit returns data quickly.

Visual Quiz - 3 Questions

Look at the execution_table, at which step does Elasticsearch load data and update caches?

A Step 1

B Step 5

C Step 3

D Step 2

💡 Hint

Refer to the 'Action Taken' column in step 3 where data is loaded and cached.

According to variable_tracker, what is the state of 'query_cache' after step 3?

A miss

B updated with query result

C empty

D N/A

💡 Hint

Check the 'query_cache' row under 'After Step 3' column.

If the request cache had a hit at step 2, what would happen next?

A Elasticsearch would load data fresh

B Elasticsearch would return cached request data immediately

C Elasticsearch would check the field data cache next

D Elasticsearch would ignore caches and execute query

💡 Hint

Look at step 2 in execution_table where a cache hit returns cached data.

Concept Snapshot

Cache management in Elasticsearch:
- Query cache stores full query results.
- Request cache stores results of filter and aggregation requests.
- Field data cache stores field values for sorting and aggregations.
- Elasticsearch checks caches in order: query, request, field data.
- Cache hits return data immediately; misses load and cache data.

Full Transcript

When Elasticsearch receives a search query, it first checks the query cache to see if the exact query result is stored. If not found, it checks the request cache, which stores results of filter and aggregation requests. If still not found, it checks the field data cache, which holds field values used for sorting and aggregations. If all caches miss, Elasticsearch loads the data fresh from disk or memory, then updates the caches with this data. This process speeds up repeated queries by returning cached data quickly. The execution table shows each step of cache checking and the actions taken. The variable tracker shows how cache states change after each step. Understanding this flow helps optimize Elasticsearch performance by leveraging caches effectively.