Login PI 3.6.23 (Sept 26, 2019)
- A public API has been made available for reading information (Web API GETs) from Login PI
- Citrix Published Apps are now supported natively. You can define one published app per environment in the environment settings
- The RDP connector has been extended with RDS Broker and RDS Gateway support (and renamed to Microsoft RDS)
- We now support storing data in an external database (Microsoft SQL). NOTE: No measurement data or configuration information is transferred to the external DB, so you are starting from scratch.
Changes / improvements:
- You can now configure data retention for “raw” data and events. After a configurable amount of days is passed, older data is removed. This “raw” data is able to be viewed in the hourly view of the charts page. The daily, weekly, monthly views and prediction are not deleted
- Notifications e-mails and events are available when a launcher went offline and when an account is disabled
- We’ve added a LogonApp to replace the logon scripts
- You can now perform system (operating system) updates of the appliance
- Nginx has been updated to improve security
- Dashboard events popups readability has been improved
- The eventlogger has been extended with a “technical events” eventlog
- Launcher version check is done on all 3 parts of the version number
- Session failures are now called engine failures
- The UI has in-line editing where applicable to make editing easier
- The UI now has pagination for improved performance and supporting a high number of items in a list
- The “Activity viewer” page has been removed
- You can now view the Login PI container status on the System page
- The “Launcher capacity exceeded” events has been renamed to "There are currently no launchers available"
- Many UI improvements
- The password configured for e-mail settings was retrieved from the back-end
- Using nested groups for AD authentication didn’t work
- In some cases the engine would fail to close an app because it was already closed previously causing the engine to crash
- Group policy measurements were incorrect in specific cases
- When using large groups of accounts or launchers they could not be selected when creating an environment due to the lack of a scrollbar
Login PI 3.5.9 (June 27, 2019)
Important notice: Because of changes on the webserver the logonscript needs to be replaced in your environment after updating the appliance. (documentation)
- A new Netscaler connector allowing you to launch ICA sessions through the Netscaler using out of the box functionality
- Support for LDAPS protocol when configuring AD Authentication
- We’ve added support for 32-bit operating systems for the launcher
- SLA Reporting is now extended with application specific results. Giving you granular information about the performance and availability of each application
- SLA Reporting now allows for customizing the logo
- You can now configure application credentials for each application. These credentials can be used in the application scripting language to login to the application being tested. These credentials are stored encrypted and retrieved by the engine using an encrypted connection
- A Login PI Eventlogger has been added. You can install this on a windows machine and it’ll create a Windows event log which will contain the events from the appliance
Changes / improvements:
- You can now change connectors in the environment configuration
- Portainer has been moved to be served by our webserver. It’s now accessible through https://applianceUrl/portainer.
- The webinterface now starts faster due to improved dependency checks on the backend
- The engine now shows a wait window on screen with a countdown when waiting in between application runs
- The Wait(); command in the engine now has the option to show a wait window on screen with a countdown and customizable message
- The ShellExecute(); command now has a parameter that configures the application to stay open after the script has ended. Up until now application started by ShellExecute would always be closed after the script ends
- The Type(); command now has a parameter that allows you to hide the Typed characters from logging. This is enabled by default. This is due to security and the content of the Type command might have sensitive data
- The web interface now shows dates and times in the locale that the browser is using
- If the engine fails to measure either logon or latency an event is generated
- The launcher version now shows up for each launcher in the launcher page and will show an indication if the launcher version and appliance version are not identical
- E-mail settings are now hidden in SLA reporting when e-mail settings aren’t configured in the External Notifications
- Initial loading time of the charting page is improved
- The webinterface now shows the release notes once after upgrading
- GPO Processing can be measured in more cases
- The custom and Horizon View connector have quotes around all fields so spaces don’t affect functionality
- SLA Reports can now be deleted
- The contents of e-mail notifications have extended with more information
- The charting page has a “Now” button allowing you to quickly navigate to the current day/time
- You can now automatically generate SLA reports on a weekly schedule
- The launcher installer now places the shortcut to the console launcher when installing on Windows 7
- Many UI improvements
- After disabling the schedule on a Horizon View environment the launcher would not kill the session after 1 minute. This would only occur when a session didn’t have an engine running
- When the engine runs while there is no environment configured 4 identical events would be generated
- The timeout on the START(); command in the engine was fixed to 45 seconds instead of the configured amount
- When disabling the schedule or multiple sessions the launcher would end all of them but the launcher UI would only update 1 of the running sessions with the “Ended” label
- If the schedule was disabled after a session was initiated but before the engine was started the engine would show an error message in the logging
- The X axis in the charting page wouldn’t start at the whole hour if there was no data present at that time
- When no measurements are selected in the “Logins” section of the charting page the chart wouldn’t show the message “Please select at least one measurement to see data”
- When a process was spawned that required admin privileges to kill the engine would fail to kill that process and logoff the session. The engine will now ignore that process
- The logoff button would show an HTTP 500 error message when using an AD account
- The login name field on the login page was labeled with the “email” tag
- In the dashboard when hovering on the login times chart the details would be mixed up. (OS would be in the protocol row for example)
- SLA Reporting page ordering was random
- When losing the connection to the backend the launcher could start another launcher process resulting in 2 or more launcher instances running on one machine
- The web server would crash when the total login time exceeded the threshold
- An account could only be edited if the password was also changed
- The scheduler is now limited to launching 20 sessions per 20 seconds. In extreme scenario’s the scheduler would “overshoot” the amount of concurrent sessions defined (100+ sessions)
- An AD account couldn’t be used to login if the group defined in AD Authentication was the primary group of the account
- If there are no launchers selected in the charting page, you cannot export login measurement to CSV
Known issues (known issue list)
- After the update the password of the admin account needs to be reset before portainer is accessible again
- Intellisense in the scripting editor only works after downgrading the C# plugin to a previous version
Login PI 3.4.12 (April 29, 2019)
- The session login steps “Display Protocol Connection”, “Profile Load”, “GPO Processing” and “Total Login Time” are now measured by Login PI. This is shown as Login Performance.
- You can now define a user group in AD Authentication that has read-only access to the dashboard, charting, events and export data
- SLA Reports are now generated in PDF format
- SLA Reports can now be e-mailed
Changes / improvements:
- The launcher installer now checks if the appliance FQDN can be resolved to an IP address. It shows a friendly error message when it isn’t.
- The launcher installer now checks for https connectivity to the appliance and if the certificate is present. It shows an error message if either doesn’t work.
- When editing an account the password field now shows asterisk character to indicate the password is already saved
- The left sidebar in the charting page now supports live updates
- Many UI improvements
- The URL used to access the webinterface can be changed from the appliance menu
- The launcher installer didn’t accept all SSL/TLS versions causing the certificate download to fail in some cases
- In the dashboard on the applications table the percentages should be colored black when 0 and no data is present (…). Red when the percentage is 1 or higher.
- When adding a duplicate account the error message would also show when the add new account dialog was closed and opened again
- In the schedule table of the environment using high numbers for the interval, concurrent sessions, wait or run applications values the interface would show a non-human readable error message and the button alignment would be incorrect
- Account groups, launcher groups, environments, applications and SLA report names could be created with identical names when using different casing
- In some cases a worker would get stuck on the authentication process
- When uploading an incorrect script the error would always be “Buffer cannot be null” instead of the specific error message with line indicator
- If the scriptworker container would restart after an incorrect script was uploaded the application would be come corrupt and no new script could be uploaded anymore
- In some cases the gateway container wouldn’t be able to reach containers after they’ve been restarted causing actions related to that microservice to fail
- If an application error contained more then 128 characters the engine couldn’t send back the error message
- The details in a popup from a application measurement in charting would be ordered randomly
- If the first measurement came in at 12:30. Charting would start the X axis at 12:30 instead of 12:00
- In some cases statistics wouldn’t be calculated so no data showed up in the daily, weekly and monthly view of the “Reporting - Charts” page
- The list of SLA reports would be ordered incorrectly
Login PI 3.3.5 (March 28, 2019)
- Active Directory integration. You can now define an AD group which is allowed administrator access to the web interface of Login PI 3
- The Citrix Storefront and Microsoft RDP connectors now allow you to define display settings (fullscreen, or a custom resolution). NOTE: To enable this feature the Launcher must be updated to version 3.3.
- Application & Login Failures are now added to the notifications section of an environment, allowing you to define e-mail notifications for those failures.
- Environment performance data export. You can now export all measurements within one or multiple environments using the “Export Data” tab located under “Reporting”
Changes / improvements:
- If the engine logs off due to a disconnected session this is specified in the “Session failure” event
- The machine name the engine was running on is added to the “Session failure” event
- The gateway container now checks if all containers are up and running when it starts, if not it shows a message that Login PI is still booting
- The data export functionality in charting is now able to filter down to measurements (previously it could only filter on applications)
- Dashboard latency average line-chart shows a balloon tip
- The red application failure indicators on the dashboard show a tooltip with additional information
- The troubleshooting page shows easier to read messages
- The troubleshooting page allows you to abort a session. If aborted, the session will logoff within 1 minute
- Disabling the schedule on an environment will close all running sessions gracefully within a minute regardless of the application configuration
- Notation of latency measurement counter is changed in the popup of the daily, weekly and monthly charts
- Environment names were all uppercase on the dashboard. Now they’re identical in casing to how they are defined in the configuration tab
- The environment dashboard counters text (in the top left, next to the environment name) has been improved for readability
- Descriptions have been added to the applications page downloads
- SLA report new thresholds are enabled by default
- SLA report list of reports are now ordered from latest to oldest by default
- We now support accounts with spaces in the account name
- The dashboard latency dial has been changed to match the other dials
- The concurrent based scheduling would launch less sessions then configured (slowly decreasing amount of sessions over time)
- PI would show a “not enough launchers” when it should show a “not enough accounts”
- If a launcher would be disconnected for a while the session counters in the launcher page would be incorrect when the launcher reconnects
- The “Launcher went offline” event timestamp was incorrect
- Visual Studio Code wouldn’t run a script with F5
- Latency wouldn’t be recorded in some scenarios if multiple protocols are used on the same machine
- The “Restart on complete” option in the desktop connector would never show an enabled status (even though it was enabled)
Login PI 3.2.8 (February 12, 2019)
- Latency measuring and (e-mail) notification for Microsoft RDP, Citrix ICA, VMware PCoIP, VMware RDP and VMware Blast
- Applications can be reordered in the environment schedule
- Measurements can be exported to CSV files from the Charts page
Changes / improvements:
- Application scripts are now compiled once per PI session instead of per application script run
- Application failures are now visualized separately from the Application Performance chart in SLA reporting
- The environment dashboards shows counters for online launchers, active sessions and account that are in use
- The engine now logs off a session when it detects the session is disconnected
- New snippets have been added to the Scripting Toolset download
- Launcher UI shows both launcher version and appliance version
- Tooltips for measurements in charting have been extended with additional information
- The SLA report Login Success chart was not filtered by environment
- The engine was reading environment variables to retrieve the environment name
- The engine was reading the incorrect environment variable for retrieving the domain FQDN the user account is in
- Some parts of application measurements weren’t deleted in the DB when deleting an application
- Orphaned sessions could be created when updating the PI appliance on the fly causing a container crash and no sessions being launched anymore
- The events on the main dashboard page contained all events ever generated instead of being filtered down to the selected timeframe (15m, 30m or 1h)
- Balloon popup doesn’t show in charting day, week and month views
- A new launcher group using a filter could only be saved when one or more launchers matched the filter
- Navigating back to the main page was impossible in some scenario’s
Login PI 3.1.42 (January 9, 2019)
- You can now generate custom events from the scripting language, that will show in the Login PI Dashboard.
- The engine automatically creates a screenshot when an error occurs in the logging folder of the session where the Login PI Engine is running.
- File and directory operations in the engine now have a continueOnError parameter
- The engine now supports reading environment variables
Changes / improvements:
- The engine now also reports back immediately after each measurement in an application script, instead of waiting until the end of the application run. It will feel much more reactive in the Login PI dashboard.
- When an environment is deleted all measurements and data related to that environment are also removed (irreversible)
- When an application is deleted all measurements and data related to that application are also removed (irreversible)
- Accounts are now ordered alphabetically
- When the dashboard percentages are below 1% it’s shows a fractional percentage
- Added a tooltip on item in the dashboard events that were cut-off by an ellipsis (…)
- Directly browsing to a URL is now possible (previously you were always redirected to the main page)
- The engine checks if the schedule is still enabled after all applications have been executed and logs off if the schedule is disabled
- Global e-mail settings frontend didn’t report back errors if you used invalid input
- The documentation tile now points to the correct link
- The engine has a new method to track processes which improves stability in closing processes
- The environment dashboard has improved sorting. It now orders on (in this specific order) Measurements over threshold, application failures and logon failures
- Environment variables can now be used in the application executable path and working directory
- When defining accounts you can also use the FQDN instead of netbios domain name. The engine supports both. (KB article needed)
- Scheduling has received a number of stability improvements
- The VMware Horizon View connector logs more info about failed connections
- The launcher issues a warning when running on windows 7. Only the console version is supported on this OS.
- Input validation has been added to the appliance menu
- Online updating is now possible from the appliance menu
- Added support for Hyper-V and XenServer (see Requirements)
- SLA Reporting would create a report for yesterday/last month when the environment didn’t exist then
- Deleting an application with results in an environment would prevent charting from loading for that environment
- Enabling the e-mail settings would already enable the feature without clicking “Done”
- Logging off action wasn’t shown in the logging when the engine fails during initialization
- Underscores weren’t properly translated to spaces in measurement names in charting
- The custom connector incorrectly reports back session failures
Login PI 3.1.34 (November 6, 2018)
This is the first generally available release of Login PI Release 3.
What's new - There are lots of new things in Release 3... but here are some highlights.
- Installation: Login PI is now delivered as a Virtual Appliance. There is no installation necessary and it only takes minutes before you are up and running.
- Logons and Applications: Logons and application scripts are now run and timed independently. You can easily turn on logons and whichever application workflows you'd like to run.
- Application Workflows: Say goodbye to the workload script that has all of the applications rolled up into one. Now applications can be added via the Web UI. For deeper application insights each application workflow can be customized in Visual Studio using our brand-new Login VSI workflow commands. Snippets are provided to make customizing super easy. Syntax highlighting, Intellisense and Autocompletion makes it even easier. All customized scripts can be uploaded using the Web UI.
- Groups and Filtering: User Accounts and Launchers are added to a general pool, enabling a higher degree of resiliency. Accounts and Launches can also be grouped to enable more discreet segmentation per environment.
- Dashboard: Multiple Environment Dashboard and deeper drill-down application performance and availability data
- Launcher: The Launcher can be downloaded from the Web UI which comes with a simple installer. The Launcher now runs as an application on the desktop
- Charts and Events
- Hour, Day, Month, Prediction Views
- Automated daily/monthly generation of SLA Reports
Differences between Release 2 and Release 3 (version 3.1.34)
- Profiles (PI2) are now called Environments (PI3)
- Thresholds (PI2) are now called Notifications (PI3)
- Connections (PI2) are now called Connectors (PI3)
- Thresholds for Notifications are not automatically generated. We advise referencing your SLAs or letting the virtual user run for a day or two and looking at the baseline values before creating a threshold.
- Target Engine: The target engine operates very similarly to Login PI 2, however a new and less obtrusive status bar will appear. Also, new application scripts are pulled every time before they run, so if you change an application script while the virtual user is logged in you will see your changes take effect on the next application run.
- Architecture: Login PI Release 3 uses Login VSI's next generation architecture. This is a modern Docker microservices architecture running on Debian, leveraging Microsoft .NET Core. Very scalable, reliable and secure.
- Logon and Latency Metrics: Login PI Release 3, version 3.1.34 will not ship with Logon statistics or Protocol Latency metrics. They will be added in subsequent versions of Release 3 very soon.
- There is no "Daily Schedule" feature in Release 3. This was sometimes used for configuring backup windows
- Migration: Login PI 2 Data cannot be migrated to Release 3. We recommend running Login PI 2 and Login PI 3 in parallel until your data retention policy has been met with Login PI 3, after which time you can decommission or backup and archive Login PI 2.
- Database: The Login PI 3 Virtual Appliance uses an embedded database, so an external database is not required or supported. Support for external databases will be provided in subsequent versions.
- Because Login PI no longer installs as a Microsoft Windows Server, Microsoft Event Viewer will no longer carry Login PI events. All events can be discovered via the ReSTful Web API.
- Login PI 3.1.34 will ship with a locally configured single administrative user. Active Directory Administrators and Read-Only users will be provided in subsequent versions.
- Active and valid Login PI 2.x licenses may be used with Login PI Release 3