Squirro 2.5.3 - Birch - Release Notes

Owned by Felix Hürlimann (Deactivated)
Last updated: Apr 03, 2018 by Oliver Seeliger (Unlicensed)
6 min read
We are super proud to release Squirro version 2.5.3!

This version was released on December 13th, 2017. The major new features are dynamic dashboards, data loader in the user interface and PDF export.

Table of Contents

Changes

New Features

  • Printable dashboards: you can now print any dashboard from your browser and it will come out looking the same way it does on your screen. Tip: switch the page orientation to landscape for best results.
  • Export as PDF beta: a much requested feature; it's now possible export your dashboards as a PDF. You can access this by enabling the "Show More button" from a Search Bar widget on your dashboard. Note there are still one or two known issues we are ironing out so we release this as a "beta" feature.
  • New Widgets
    • Enrich your Dashboards with the Rich Text widget allows to place formatted text on your dashboard, including images and links to other content.
    • See the World with the Region Map widget, that allows you to plot data on maps or world or selected countries, clustering that data by regions
    • Make Numbers Stand Out with the Metric Widget, which allows a simple way to place a key metric on a dashboard, such as a percentage or a counter.
    • Time Travel through your your data with the Calendar Widget, which provides an alternative to the timeline when you need to explicitly pick dates.
    • Keep your place by enabling the new Breadcrumb mode on the Reset widget and you'll see Breadcrumb Navigation of which filters you've applied as your drilled down into your data.
  • Dynamic Dashboards: we added dashboard layers in version 2.4.5 as a feature for developers. Now we've added them to the Dashboard Editor so everyone creating dashboard can take advantage. You can switch layers in your dashboard based on conditions such as where a facet is selected or whether the search contains a certain result. By switching layers you can show completely different content, allowing you visualise results in the way that fits best to the data.
    CHANGE The dashboard layers replace the concept of dashboard loaders which are no longer supported. Please see Breaking Changes below.
  • Everyone can load data from everywhere: the Squirro Data Loader has long been the tool for developers to load data from sources anywhere outside or inside your company and via the use of Data Loader Plugins it's easy to extend it to any new source. With this version of Squirro we're extending this power to everyone creating Squirro projects with what's known as the Self Service Dataloader. The Self Service Dataloader allows you to publish Data Loader Plugins directly into the Squirro "Load" tab and provide a simple point-and-click experience to configure the plugins. We've also taken the opportunity to replace our old CSV data loading feature with a two new new ones based on Self Service Dataloader architecture, making it painless to import CSV and Excel files into Squirro.
  • Simple API for Recommendations: for developers you can now get recommendations for content related to a given item id or a string of text you provide to the API - see the API documentation for details.
  • Easier Clusters on RHEL7: Added support for Multi-node clusters in squirro-cluster-nodes on RHEL7 / Centos 7
  • Analyse Performance Bottlenecks: we've added support for using Jaeger to analyse performance throughout Squirro clusters and identify bottlenecks. To use Jaeger on your cluster see the Performance investigation.
  • Help us help you with performance issues: we've also introduced New Relic Browser Monitoring which allows us to analyse frontend and dashboard performance issues. This is enabled by default, which means frontend performance statistics are make available to Squirro's engineering team via New Relic's cloud. If you wish to disable it this can be done with a simple configuration change - see configuration.ini for details.
  • Weighted Tags alpha: to support our Trinity strategy of bringing pragmatic AI to Squirro, we've extended our tagging featuring to allow percentage weights to be attached. This makes it possible to attach a measure of certainty to a tag on an item, enabling - for example - items to be a assigned to weighted clusters by a trained ML pipelet. This feature is currently alpha level - if you'd like to start using it please contact your partner representative
  • SAML Support: via Squirro Studio under the Server → Settings menu, we've added support SAML making it possible for users to login to Squirro via identity providers such as Microsoft ADFS. See SAML Single Sign-On for Squirro for full details.
  • Usage Analytics: Squirro can low log user activity, so it's possible to know how many people are looking at dashboards for example. It is disabled by default but can be enabled by following the instructions here.

Improvements

  • Significant Performance Improvements: with help from Jaeger and New Relic we've been able to significantly improve performance in a number of areas, including how 
  • Tuning Type-Ahead: we tweaked the way type-ahead works to make it even easier and friendler to us, such as applying ellipsis to typeahead matches where the match string is long.
  • Squirro RPMs now depend on python27 version
  • Performance Tracking of every single User Query (added on December 14, 2017 with release 2.5.3-4098)

Bug Fixes

  • Browser Specific Issues Fixed
    • Firefox: Settings tab now has correct padding
    • IE11: toast's text is now vertically aligned
  • Smart Filters Issues Fixed
    • Smart Filter Editing: fixed issue where clicking done overlay closes modal
    • Fixed styling of smartfilter advanced table headers
    • The "Save Smart Filter As..." dialog label now turns blue when you select the input field
    • Smart Filter Create / Edit: Tag cloud is now rendered
  • Widgets Issues Fixed
    • Cleaned up word cloud when there are many words
    • Made it possible to click on words in the Significant Terms Widget
    • Respecting additional query in Heat Map
    • Facets Table widget now able to scroll item detail
    • Facet table now works when one of the columns is set to Body
    • Time selection widget now shows current state
    • Proper position handling in custom widgets (added on December 14, 2017 with release 2.5.3-4098)
  • Miscellaneous Issues Fixed
    • Removed flicker when requesting API access token
    • Fixed width of item links on mobile
    • Fixed issue where facets visibility checkbox fails
    • Dropdown for projects is now correctly positioned
    • Fixed dataloader for newsriver
    • Prevented resume_source from deleting title
    • Facet settings page filter now shows non-analyzed string
    • Fixed sorting on keyword field
    • Dashboard Edit: text area sizes are now dynamic
    • Proper facet type checking (added on December 14, 2017 with release 2.5.3-4098)
    • Hardened Squirro Client ItemUploader to properly handle empty item list (added on December 14, 2017 with release 2.5.3-4098)
    • Fixed exporting of dashboards as PDFs (added on December 14, 2017 with release 2.5.3-4098)
    • Removed query prefix ("removed" on December 14, 2017 with release 2.5.3-4098)
    • Fixed snapshot generation after you remove the query from the searchbar (added on December 14, 2017 with release 2.5.3-4098)
    • Better tracking of Pipeline 2.0 data ingestion statistics (added on December 22, 2017 with release 2.5.3-4101)
    • Dashboard display fixed for items with no near-duplicate items (added on December 21 and 22, 2017 with releases 2.5.3-4099 and 2.5.3-4101)
    • Fixed an issue by which the pipeline 2.0 "opt-in" was opting in to additional builtin enrichments that hadn't been opted in to (added on January 15, 2018 with release 2.5.3-4102)
    • Fixed installation of the squirro-sql-config package on systems not necessarily running bash as a shell, also fixed erasing of individual Squirro services (added on January 17, 2018 with release 2.5.3-4103)
    • Fixed the pipeline 2.0 to work properly with pipelets configured via the squirro_data_load tool, e.g. --source-profile (added on February 5, 2018 with release 2.5.3-4104)
    • Fixed the filtering enrichment step to work with date-like values in keywords (added on March 1, 2018 with release 2.5.3-4108)
    • Pre 2.6.0 fix to not assume that services such as pipeline 1.0 are still running upon removal. This is useful to aid in the replacement of pipeline 1.0 with pipeline 2.0 (added on April 3, 2018 with release 2.5.3-4113)

Breaking Changes

With the release of the layers, the dashboard loader api has changed.
Loaders can no longer modify the dashboard sections.

Affected functions are:

  • this.dashboard.sections.reset()
  • this.dashboard.sections.add()

If you are using these functions in loaders, then please follow the process described at Upgrading from Dashboard Loaders to Layers.

Installation Instructions

Fresh Installation

Please follow the regular installation steps

Upgrade

Please ensure that your current version is 2.5.2. If you are on a version older than 2.5.2, please contact support.

Additionally if you are using Squirro in a Box, additional steps are involved. In this case we also ask you to contact support.

Storage Nodes and Cluster Nodes on the same machine

If the storage and cluster nodes are installed on the same machine / VM, then following this process:

# ensure that you have version = 2.5.2 installed, otherwise please update to version 2.5.2 first.
sudo su
cat /etc/squirro/version.ini
yum update squirro-storage-node-users
yum update squirro-storage-node
yum update squirro-cluster-node-users
yum update squirro-*

Storage Nodes and Cluster Nodes on separate machines

If the storage and cluster nodes are installed on separate machines / VMs, then following this process.

  1. Update all storage nodes with these commands:

    # ensure that you have elasticsearch = 5.6.0 installed before proceeding, otherwise please update to version 2.5.2 first.
curl -XGET localhost:9200/
sudo su
yum update squirro-storage-node-users
yum update squirro-storage-node

  2. After having updated all storage nodes, update all cluster nodes with these commands:

    # ensure that you have version = 2.5.2 installed, otherwise please update to version 2.5.2 first.
sudo su
cat /etc/squirro/version.ini
yum update squirro-cluster-node-users
yum update squirro-*