Chat Support
Loupe - Log - Monitor - Resolve
Loupe / User's Guide / Loupe Server / Loupe Self-Hosted / Loupe Server Administration - Elasticsearch Requirements
In This Topic
    Loupe Server Administration - Elasticsearch Requirements
    In This Topic

    Loupe can use Elasticsearch as its full text search engine and for Loupe Monitor.  This is the recommended configuration for Loupe Server Enterprise, particularly larger installations, as the built-in full text index system (Lucene.net) is not intended to work against network file systems.

    Elasticsearch Requirements

    Loupe requires Elasticsearch 7.0 and later, configured with either no authentication (the default, but not recommended for production, configuration) or with basic authentication.

    Elasticsearch should not be run on the same hosts as Loupe itself as it requires significant memory and processor to run. 

    Elasticsearch Capacity Planning

    Loupe will store a copy of every Event and Issue in the full text search system.  This is effectively a copy of each unique warning and error in each session.  Unlike the main database this data can't be compressed within Elasticsearch as it has to be indexed.  As a general rule of thumb this takes five times the equivalent SQL storage capacity.

    If Loupe Monitor is configured for an application, the detailed log information is stored in Elasticsearch.  This is space limited up to the session data limit for the repository.

    Recommended Configuration

    Since Loupe does not store any information exclusively in Elasticsearch it isn't necessary to have a redundant set of data nodes to avoid data loss.  If the indexes are not available in Elasticsearch Loupe will recreate and repopulate it.  This may take time, but there is no loss of data. 

    Most installations can function well with a single Elasticsearch node but it's also possible to create a multi-node cluster for greater capacity in large installations.  The key recommendation is to have an odd number of master nodes to ensure the cluster is managed correctly.