Time-To-Live (TTL) indexes

TTL (Time-To-Live) indexes are useful when you need documents to be deleted automatically after a certain time. It's ideal for temporary records, logs, or session data that you don't want to keep indefinitely. Unlike most indexes that focus on query performance, TTL indexes serve a different purpose by ensuring outdated data is deleted periodically and efficiently.

They are single-field indexes and not allowed as multi-field/compound indexes.

note

• Documents inserted before the index was created may not be affected until an insert or update operation occurs. It is recommended to create the TTL index before inserting documents.
• The cleanup of expired documents is a background operation that runs every 60 seconds, so deletion may not be immediate.

Creating a TTL index

A TTL index is created using the createIndex command with the field you want to index and the expireAfterSeconds option set to the number of seconds before a document expires. The field must have a Date type.

For example, let's create a TTL index for the reservation.date field in a books collection.

db.runCommand({
  createIndexes: 'books',
  indexes: [
    {
      key: { 'reservation.date': 1 },
      name: 'reservation_ttl',
      expireAfterSeconds: 60
    }
  ]
})

Let's insert a document into the books collection:

db.runCommand({
  insert: 'books',
  documents: [
    {
      title: 'The Great Gatsby',
      author: 'F. Scott Fitzgerald',
      reservation: {
        user: 'Ethan Smith',
        date: ISODate('2025-03-15T11:00:00Z')
      }
    }
  ]
})

After 60 seconds, the document will be deleted in the next cleanup operation.