ElasticSearch RAM requirement


We are thinking of using ElasticSearch to store log data with the following simple schema:

    "data" : {
        "properties" : {
            "date": {
              "type" : "date",
              "format" : "yyyy-MM-dd HH:mm:ss"
          "status": {"type" : "string"},
          "group": {"type" : "string"},
          "ip": {"type" : "string"},
          "username":{"type" : "string"},
          "category":{"type" : "string"},
          "url":{"type" : "string"}

We are estimating that we will have around a maximum of 5000 INSERT/s. Any idea if a 8GB of RAM will be enough to handle this?
From what I can see on the official documentation, the typical server uses between 16-64GB but no information about INSERT/s.

Thanks for your suggestions.

While insertions per second affect the RAM requirements, I don't think it's that important. The total amount of data and the query patterns are way more important. So, I'm sure a 5000 msg/s inbound rate will be fine for a while, but eventually you'll have so much data that an 8 GB heap won't be enough.

1 Like