MongoDB: Schema for storing read receipts of messages in a chat system

Question

We are implementing read receipts in a chat system. We want to store the timestamp when each message was read by the participants. (i.e messageId, participantId and timestamp), and display this when the user wants to see message info on the frontend. Other than this we also want to display ticks next to each message, depending on whether all participants have seen the message.

We are using sockets for sending the event when the message(s) is seen.

As in a chat system, the number of messages will keep growing, and we are concerned about the scalability of the system.

We have thought of the following schemas/approaches as of now:

1. Creating entries for the participant and the message in the 'ReadMessages' collection when a message is seen. When fetching the messages, it has to be populated with ‘ReadMessages'.

2. Embedding this data in the messages collection itself. This approach would require frequent updates to the 'Messages' collection, and probably slowing down reads on this collection?

3. Creating entries in the 'ReadMessages' collection when a message is seen by a participant, and storing a flag named 'Read' in the 'Messages' collection, that will be re-calculated and updated (if required), whenever a message is seen. This approach doesn't require populating while fetching messages, and also minimizes updates.

4. Similar as 3. Just the status 'Read' in the messages collection will be updated in a cron job. This will require lesser processing when a message seen event occurs, but this cron job will have to scan through all messages to see which are not marked as 'Read' yet (as the index is only on _id). Also, read receipts on the client will be updated with a delay depending upon the interval in which the cron job is run.

Which of these would be better in the long run? Or is there any other better way schema for this can be designed?

Also, could multiple concurrent inserts to the collection cause performance problems?