Mongo has a feature called the oplog, where inserts/updates/deletes are recorded for replication purposes (so that another Mongo node can read these operations and keep itself up-to-date). The oplog can be leveraged by other systems to do things like keep ElasticSearch in sync with Mongo.


Existing implementations of oplog tailing are hard to get started with and/or are not performant enough. We wanted to a better solution.

How Far Did We Get?

  • Cleaned up the existing code and made it very easy to get a working oplog tailer up and running
  • Code is now open source on Github.

What’s Next?

  • Get the jar on Maven Central
  • Use it for other things:
    • Smarter handling of indexing (instead of polluting the code with extra calls, we can watch the db for changes)
    • Performance/db operations auditing