Expand all | Collapse all

How to Interpret Memory Usage in Admin Console

  • 1.  How to Interpret Memory Usage in Admin Console

    Posted 05-25-2017 17:33
    A memory usage window in the admin console always has two graphs. Which one is the actual usage and what is the function of the other graph?

    For instance, in the image included below, there is one graph that hovers around 7GB and one that is around 1.5GB. Which one represents the actual usage and what does the other value represent?

    In the following image, there you can see that some values have changed:

    Cheers - Geert

  • 2.  RE: How to Interpret Memory Usage in Admin Console

    Posted 05-26-2017 12:10
    Great question Geert!

    The first graph in the admin console represents the Elasticsearch container. Elasticsearch is responsible for indexing related clusters of data nodes to work together and respond to requests. We currently have 1 index per tenant, where the index is broken up into smaller segments in order to increase search efficiency. Elasticsearch is used abundantly in Jama and is expected to use more memory than most of our other containers.
    The second graph in the admin console represents the Nginx container. Nginx is responsible for acting as a load balancing layer. It has a lightweight infrastructure because it it process driven, rather than event driven. Therefore, Nginx defines worker processes to serve requests, instead of spawning a new process for each request. That is why the memory usage represented on the Nginx graph is significantly lower than the Elasticsearch graph.

    Hopefully this answered your question!

    Jamal Adams
    Jama Software
    Portland OR

  • 3.  RE: How to Interpret Memory Usage in Admin Console

    Posted 05-29-2017 18:44
    Hi Jamal,

    I am sorry, but this answer doesn't address my question at all.

    My question is about the two lines that I see in every image. What does each line represent?

    Cheers - Geert

  • 4.  RE: How to Interpret Memory Usage in Admin Console

    Posted 06-01-2017 11:29
    Edited by Jamal Adams 06-01-2017 11:33
      |   view attached
    My apologies Geert.
    The original intent of the memory usage graphs were to provide data for actual memory usage, swap, and cache. However, Replicated found the graphs unclear and recently updated the graph display of this information. In recent 2.x release, Replicated has combined the data from all of the containers that are currently using memory into one graph. It also provides information on what container each line on the graph corresponds to when you hover over the data with your mouse. The attached image shows the most recent graph display.

    Specifically to the Nginx graph that you provided, the line with the changing values represent the actual memory used. It indicates that a process was freed up in memory in the container.