Skip to content
Permalink
Branch: master
Commits on Jun 24, 2019
  1. Merge pull request #1601 from alphagov/fix-load-search-analytics

    bilbof committed Jun 24, 2019
    Fix search-analytics-data-load
  2. Fix search-analytics-data-load

    bilbof committed Jun 24, 2019
    This fixes a bug introduced in 3eacc88.
    
    Setting StringIO.pos = 0 is forbidden and results in a
    error: Errno::ESPIPE: Illegal seek @ rb_io_set_pos - <STDIN>
    
    Using rewind is also not permitted, since we are reading from data
    that is being piped in.
    
    The workaround is to store the stream as a temp file, this seems
    like the nicest way to do it! This lets us rewind.
    
    Trello: https://trello.com/c/Q4nK0icC/828
  3. Merge pull request #1598 from alphagov/fix-migrated_indices-query

    bilbof committed Jun 24, 2019
    Make FormatMigrator fetch migrated_indices from specified cluster
  4. Make FormatMigrator fetch migrated_indices from specified cluster

    bilbof committed Jun 24, 2019
    Currently FormatMigrator fetches migrated_indices index
    names from the default cluster. Now that we want results
    from specified cluster, FormatMigrator needs to know that
    it should get the index names for the correct cluster.
    
    Trello: https://trello.com/c/9bjDVKo6/824
Commits on Jun 21, 2019
  1. Merge pull request #1593 from alphagov/cluster-page-traffic

    bilbof committed Jun 21, 2019
    Load page traffic into all active clusters
Commits on Jun 20, 2019
  1. Load page traffic into all active clusters

    bilbof committed Jun 19, 2019
    This loads page traffic data into all indexes,
    not just the primary one.
    
    There's possibly better ways to do this (e.g.
    copying from one index to another?). But this
    is consistent with the way data enters into all clusters.
    
    One thing that might be nice is to create a worker
    for each cluster.
Commits on Jun 12, 2019
  1. Merge pull request #1580 from alphagov/healthcheck-clusters

    bilbof committed Jun 12, 2019
    Healthcheck all clusters
  2. Healthcheck all clusters

    bilbof committed Jun 11, 2019
    Now search-api talks to multiple clusters, this will
    ensure that we healthcheck the connectivity to all
    clusters, not just the default one.
    
    Trello: https://trello.com/c/SF4gxCrs/797-healthcheck-for-elasticsearch6-connectivity-s
  3. Merge pull request #1569 from alphagov/select_cluster

    bilbof committed Jun 12, 2019
    Allow Search API to talk to multiple elasticsearch clusters
  4. Allow Search API to talk to multiple elasticsearch clusters

    bilbof committed Jun 4, 2019
    This enables search-api to send requests to multiple
    elasticsearch clusters. You can specify the clusters
    you would like to query using the query param cluster.
    
    E.g. /search.json?cluster=B will query elasticsearch6.
    The default behaviour is to query cluster A, which is
    currently elasticsearch5.
    
    This is needed now because we want to upgrade from
    elasticsearch version 5 to 6, by querying 5 or 6 ad hoc.
    
    This will have the side effect that we will be able to
    do a/b tests and compare indexes/clusters more easily.
    
    All write actions on clusters should now be performed
    on all active clusters in an environment.
    
    Some rake tasks now have an optional cluster argument,
    but by default rake tasks will run against all clusters.
    
    Trello: https://trello.com/c/uTsa3fpt/745.
You can’t perform that action at this time.