v5.1.0

Important Note

• Due to backend dependency updates it is advised to change the driverClassName in your openSpeedMonitorConfig.yml to com.mysql.cj.jdbc.Driver.
• In case you are getting issues with your mysql database and UTC timezone please refer to this: https://stackoverflow.com/questions/26515700/mysql-jdbc-driver-5-1-33-time-zone-issue

Features

• The new homepage shows all active applications with their total customer satisfaction index as well as failing measurement jobs.
• Graphite Server integration
  • Transfer information about the health status of your jobs to graphite.
  • Add new graphite servers during the configuration process.
• New information card on the application dashboard.
  • It shows how many jobs are currently failing.
  • It helps navigating directly to the failing jobs.
  • It allows to configure graphite server integration.
• The following new metrics are available:
  • Hero timings.
  • The size of javascript, image, css and html files.
  • First meaningful paint.
  • First contentful paint.
• Percentile replaces median
  • Median-button on barcharts got replaced with percentile button coupled with a slider to choose the percentile.

Improvements

• Duplicated page templates were merged into a single, better maintainable one.
• Failed job results won't be persisted anymore to reduce unnecessary data.
• The queue dashboard is now asynchronous which decreases initial load time.
• Unknown browsers and locations will be added automatically to avoid unusable incoming results.
• Failing jobs will be highlighted in the job list for a quicker overview.
• For every job schedule a corresponding job result will be created even if the test could not be started at all. This helps realizing problems communicating due to unreachable WebPagetest servers.

Bugfixes

• Fixed incorrect CSI Benchmark sorting when using different filter.
• Button for CSI Weights upload has an icon now.
• Fixed broken prompt showing up when PageMapping was removed.
• Fixed too long information text on threshold tab.
• Fixed too short space between buttons on the threshold editing page.
• Fixed bug that removes the selected threshold but also unsaved thresholds when deleting threshold.
• Fixed wrong button state when new measured event gets created.
• Fixed bug that changes the selected application when a different language option gets set.
• Fixed sticky table header error in script table.
• Fixed CSI circle turning grey when selected application gets changed.
• Fixed bug that removes the user role when a user property gets changed.
• Fixed the CRUD-views not showing the corresponding toolbar.
• Fixed job results so they can be saved without the job having a label.
• Fixed missing navigation between result pages.
• Fixed duplicative toolbar in some CRUD-views.
• Fixed bug that prevents creating or deleting users.
• Fixed bug that prevents editing graphite server.
• Fixed nightly clean up not removing references to old event results.
• Fixed losing role when changing the user name or changing its role.

v5.0.1

Hotfixes

• Fixed a bug that caused threads to stop working when persisting into detail analysis service failed
• Fixed distribution chart not being drawn
• Implemented a workaround for the calculation of the CSI for pages on the application dashboard, caused due to a race condition between CSI for jobGroups and for pages.

v5.0.0

Release Notes v5.0

Features

• Introduces the new Application Dashboard to show the performance of your application at a look.
  • It shows the current CSI of your application (JobGroup) and an overview of the last 4 weeks.
  • It shows all measured pages of your application with the current CSI and the metrics Document Complete, Speed Index and Bytes Fully Loaded.
  • It helps you to find broken measurements and CSI calculations by showing outdated CSI values and missing measurements.
  • It helps you to create and configure your CSI for your application.
• Travis CI Integration
  • Run unit, integration, geb, jasmine and angular tests on Travis.
  • Build and push docker image to Docker Hub (Deployment to our instances is still done through Bamboo)
  • Check out our build status live in the ReadMe and for details directly on Travis

Improvements

• Upgraded to Grails v3.3.6. This includes Spring 1.5.14 and Tomcat 8.5.31. This fixes CVE-2005-2090 which affected previous versions and was reported to us by cegue and irivera007
• Upgraged FontAwesome to v5.1.
• Upgraded jQuery to v3.3.
• Upgraded GORM to 6.1, forcing us to revise database flushing. Please post issues if you are experiencing problems.
• Deprecated GORM transaction annotations were substituted.
• Introduced Angular for building components and sub pages.
• Removed Vue.js and replaced the Vue.js components with Angular components.
• A contribution markdown with new development guidelines was added.
• Removed sensitive data from the logging.
• Undefined pages won't be included in the JobGroup aggregation.
• Jobs now always use the "Capture DevTools timeline" feature.
• The Docker image doesn't rely on a specific user id anymore, which makes it usable in environments like OpenShift that run the image with arbitrary UIDs.

Bugfixes

• Fixed the parsing of the scripts with variables. The pages were always undefined.
• Fixed wrong sized, wrong positioned and broken waiting spinners in charts.
• Fixed broken URL of the PageComparisonChart. The selected dates were wrong
  after loading the URL.
• Fixed deleting an empty comparison in PageComparisonChart not enabling the show button again.
• Fixed PageAggregationChart showing improvement and deterioration on one bar at the same time, due to switching between median and average.
• Fixed PageAggregationChart throwing internal server error when selecting broken results.
• Fixed filter in the JobGroupAggregationChart not working properly
• Fixed Javascript error in the header of the job list. The sticky-table-header function is not available anymore.
• Fixed type error in time series graph when trying to add a graph alias.
• Fixed the filter of the JobGroup list. After a few changes of the filter some of the jobs were not shown anymore.
• Fixed the CRUD-View of the OSM configuration. The default configuration of the OSM was loaded from the database instead of the own configuration.
• Fixed adding page mapping to CSI redirecting to page 404. The OSM was loaded from the database instead of the own configuration.
• Removed animation/transition when switching between median and average, since this led to wrong bar behaviour.
• Removed button to delete CSI-Configuration in JobGroup view. This led to a JavaScript error and was overall confusing, since you deleted the config for all JobGroups.
• Workaround: Login will now always redirect to landing page after a successful login. This prevents the appearance of 404 errors.
• Fixed the deletion of custom time series dashboards.

Code

• WPTProxyController was removed. The controller is not needed anymore.
• Median and average calculation have been moved to the QueryBuilder.
• WPT-Status-Code are now enums and are not used directly in code.

Although highly tested by us, we advise you to backup your database before upgrading.

v4.6.0

Features

• You can switch between displaying the average and the median in various charts like Page Aggregation, Job Group Aggregation, and Page Comparison.

Improvements

• Improved performance of the Time Series chart.
• We got rid of bower in favor of using only npm.
• Fixed two potential security vulnerabilities (potential cross-site injections in IE, outdated dependencies).
  Many thanks to @cegue and @irivera007 for their hints and support regarding these issues.
• Sizes on Time Series dashboard are displayed in Megabytes, not Kilobytes.

Bugfixes

• Logging works again. Our logback configuration was overwritten by another one that shipped with an outdated swagger dependency.
• Scripts can be deleted if they are not referenced in any jobs.
• Several buttons were clickable even if when being disabled.
• Fixed a problem with german locale when trimming results on the Time Series chart.
• The job status indicator showed "no data" when all 150 last job executions failed.
• Swagger UI can now be used for testing the requests (didn't work before).
• Fixed the version of mysql to 5.7 in the docker-compose files. Migration to mysql 8.0 doesn't work, yet.
• The bars in the Page Comparison chart were overlapping in the wrong order.
• CSV download of Time Series chart works again.
• PNG download of Page Comparison missed icons in the legend.

Code

• Major refactoring of integration tests for better maintainability.

v4.5.2

Bugfixes

• While creating a new job it wasn't possible to create a new job group.
• Fixed a positioning offset in the page aggregation bar chart when comparative time frame was selected.

v4.5.1

Bugfixes

• Selected items in dropdowns didn't get displayed correctly, due to an Issue in bootstrap-chosen
• Some highlighting in measurement scripts was broken.
• A minor issue in the REST API interceptor resulted in a permission problem.

v.4.5.0

Features

• We introduced a feature to create thresholds (lower and upper boundaries for a specified measurand) in a job. Based on those threshold a script can be exported. The script runs an OSM job, waits for the result and evaluates the performance metrics based on your defined thresholds. You can hook up the script in your CI pipeline and get an instant response, rating your web performance.
• It is now possible to send health metrics of the detail data service (DDS) to Graphite.

Improvements

• Replaced our REST API documentation with Swagger UI.
• Internal refactoring of the page comparison chart.
• Refactoring of PNG download mechanism enables dynamic resizing of images in a modal dialog.
• Integrated vue.js.
• Detail data service (DDS)
  • DDS uses OSM domains as identifier instead of full URL now.
  • DDS drops all OSM API keys and creates configured keys on each startup now.
  • Performance optimization.
  • We build a nightly cleanup. All detail data older 60 days is deleted per default. This duration can be configured.
  • The number of threads for downloading wpt results and the maximum number of fetch jobs to queue can be configured now.

Bugfixes

• Bar charts didn't get displayed correctly in Firefox.
• Copying a job's script to clipboard works now again.
• When sorting bars in page aggregation chart they disappeared sometimes.
• Fixed list of failed measurements.
• Fixed some problems when selecting user timings as measurand.
• Query parameters got cut off in a navigate when sent to WPT server (see Issue #31).
• Chrome tab in job settings was sometimes hidden (see Issue #32).
• Editing the Graphite server url threw an exception.
• Event results of undefined pages were not sent to Graphite.
• Active loading of detail data produced three duplicate copies of the same detail data set.
• Passive loading of detail data in the background was broken.

v4.4.0

Features

• Measurement and visualization of User Timings set within measured pages are fully supported now. The User Timings are listed and shown as additional measurands in all dashboards if measurements contain arbitrary user timings.

Improvements

• Improved performance of Distribution Chart.
• Only show scorebar on chart dashboards for relevant measurands (load times).
• Internal refactoring of the JobGroup aggregation chart.

Bugfixes

• Fixed a bug causing detail analysis dashboard to break when opening table with many assets multiple times.
• Redirect to Job list after executing Job actions like executing or deleting a job to avoid repeating action execution on page reloads in the browser.
• Fixed Js-NaN exceptions being thrown when resizing the browser window while showing a page aggregation chart.
• Fixed backend exceptions when calculating normalized values leading to no chart being drawn.
• Fixed connectivity option native to be called job/nextExecution in UI.
• Fixed RuntimeException when switching from joblist to timeseries graph with the graph button.
• Disable stacked bars for comparative time frames in page aggregation.

v.4.3.3

Improvements

• Error pages don't contain any detailed informations about the specific error anymore. Previously this was just the case for some error response codes.

Bugfixes

• Increased maximum allowed length of webpagest location names. It was limited to 50 characters and some locations of webpagetest.org server exceeded this length so they can't get fetched anymore.

v4.3.2

Bugfixes

• No aggregated Customer Satisfaction Index metrics got calculated if CSI Systems were present.
• Enabled deletion of Locations and Browsers via All Controllers menu entry.
• Improved displayed message, if Locations, Browsers or Scripts can't be deleted because associated domains are referencing them.
• Removed remains of old domain class AggregatorType.
• Fixed displaying native connectivity with wrong label.