Point-in-time API in Elasticsearch - Step-by-Step Execution

Concept Flow - Point-in-time API

Start Search Request

↓

Open PIT (Point-in-Time)

↓

Use PIT ID in Search

↓

Fetch Consistent Snapshot

↓

Return Search Results

↓

Close PIT (optional)

↓

End

The Point-in-time API opens a snapshot of data to search consistently, then uses that snapshot ID in search requests to get stable results.

Execution Sample

Elasticsearch

POST /my-index/_search
{
  "pit": {"id": "PIT_ID", "keep_alive": "1m"},
  "query": {"match_all": {}}
}

This code uses a PIT ID to search a consistent snapshot of 'my-index' for 1 minute.

Execution Table

Step	Action	PIT ID	Search Query	Result Snapshot
1	Open PIT on 'my-index'	abc123	N/A	Snapshot created
2	Search using PIT ID 'abc123'	abc123	{"match_all":{}}	Consistent results from snapshot
3	Search again with same PIT ID	abc123	{"match_all":{}}	Same consistent snapshot results
4	Close PIT 'abc123'	N/A	N/A	Snapshot closed
5	Search with closed PIT	abc123	{"match_all":{}}	Error: PIT not found

💡 PIT closed or expired, no longer available for consistent search

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5
PIT ID	None	abc123	abc123	abc123	None	None
Search Result	None	None	Snapshot data	Snapshot data	Snapshot data	Error

Key Moments - 2 Insights

Why do we need to use the PIT ID in every search request?

Because the PIT ID points to a fixed snapshot of data, using it ensures all searches see the same data state, as shown in steps 2 and 3 of the execution_table.

What happens if we try to search with a PIT ID after closing it?

The search fails with an error because the snapshot no longer exists, as shown in step 5 of the execution_table.

Visual Quiz - 3 Questions

Look at the execution_table, what is the PIT ID after step 1?

A None

B abc123

C def456

D Error

💡 Hint

Check the 'PIT ID' column in row for step 1 in execution_table

At which step does the search result become an error?

A Step 5

B Step 4

C Step 3

D Step 2

💡 Hint

Look at the 'Search Result' column in execution_table rows

If we do not close the PIT, what happens to the PIT ID after step 4?

A It causes an error on next search

B It becomes None immediately

C It remains valid for further searches

D It changes to a new ID

💡 Hint

Refer to variable_tracker for PIT ID after step 4 and the exit_note

Concept Snapshot

Point-in-time API lets you open a snapshot of your data for consistent searches.
Use the PIT ID in each search to see the same data state.
Close the PIT when done to free resources.
Without PIT, searches may see changing data.
PIT IDs expire or close, causing errors if used afterward.

Full Transcript

The Point-in-time API in Elasticsearch allows you to open a snapshot of your index data to perform consistent searches. First, you open a PIT which returns a PIT ID representing a fixed snapshot. Then, you use this PIT ID in your search requests to ensure all results come from the same data state, even if the index changes. You can perform multiple searches with the same PIT ID to page through results or repeat queries consistently. When finished, you close the PIT to release resources. If you try to search with a closed or expired PIT ID, you get an error. This process helps avoid inconsistent search results caused by data changes during querying.