Changelog
2025.12
Playwright Scripting
Playwright Test is now generally available to everyone for load testing and monitoring!
You might choose Playwright Test cases as an alternative to Loadster’s own Browser Bot scripts if you already have a Playwright Test suite of test cases, have Playwright and Playwright Test experience, or simply prefer write your scripts in a text editor with an open source testing framework.
Here’s an example of a simple Playwright Test script:
const { test, expect } = require('@playwright/test');
test('Register Account', async ({ page }) => {
await page.goto('https://petstore.loadster.com');
// Validate the page title
await expect(page).toHaveTitle(/Dangerous Pets/);
// Fill in a username and password
await page.getByTestId('register-username').fill('user1');
await page.getByTestId('register-password').fill('s3cr3+');
// Submit the form
await page.getByTestId('register-form').locator('button').click();
});
You can use these Playwright Test scripts just like other Loadster scripts, both in load tests and site monitors.
Read more about Playwright Scripts in Loadster and give it a try!
Monitoring Performance Thresholds
Loadster monitors can now evaluate performance-related thresholds for each monitor cycle as pass/fail criteria.
You can set monitoring performance thresholds based on Response Time (Average), Response Time (Total), and Cycle Duration.
If a monitor cycle falls outside these thresholds (below the minimum or above the maximum) the cycle is treated as a failure even if no other errors were encountered.
Other Improvements & Bug Fixes
- Fixed a problem with the Incidents list affecting customers with many past incidents.
- Fixed a bug with showing/hiding individual bot groups on a finished test report.
- Reduced the time to start playing scripts from the editor.
- Improved the name filter on the project name to be “sticky” and remember past searches after navigating away.
- Optimized page load time of the main dashboard and project pages.
- Fixed a double-paste issue when pasting into a field in the script editor.
- Reduced cold start time when playing scripts on AWS Lambda.
- Fixed a race condition that occasionally resulted in a blank report (requiring a page refresh).
- Made past script runs appear in the history when returning to a script that was played previously.
- Fixed a bug with file uploads in scripts when running on shared engines.
- Improved colors and formatting in notification emails.
2025.11
Improvements & Bug Fixes
- Fixed a stale state bug that caused the top nav test indicator to keep spinning when a test was no longer starting.
- Fixed the Console tab in the Trace dialog so it shows up consistently.
- Fixed minor UI formatting issues in the test report.
- Improved navigation of the script Timeline and corresponding modal dialogs.
- Improved usability when enabling or disabling monitors.
- Improved usability of the script run history component for viewing previous script runs.
- Improved usability around copy and paste of steps in the script editors.
- Added more international fonts to the Chrome installation used by Browser Bots.
- Optimized performance of the home dashboard and monitor summary components.
Further Reading
The addition of international fonts to the Chrome installation used by Browser Bots is worth highlighting. If you’re testing websites that render text in Chinese, Japanese, Korean, Arabic, or other non-Latin scripts, the browsers need the right fonts installed to display those pages correctly. Without them, you might see blank squares or garbled text in screenshots, which makes debugging harder. This update ensures Browser Bots can render a wider range of languages accurately.
The Site & API Monitoring performance optimizations also help keep the monitoring dashboard snappy, especially if you have a lot of active monitors.
2025.10
Flexible Load Stages
You can now build scenarios with multiple ramps and peaks, not just a simple ramp up and down. Adjust the ramp pattern, target bots, and duration of each stage. This is helpful for spike testing, starting and stopping different bot groups at different times, and other interesting ramp patterns.
Live Load Adjustment
While a load test is running, you can now edit the ramp patterns in real time. You might want to do this if the test has reached a spike or peak that’s causing errors against your site, and you want to continue your load test but with fewer bots than initially planned. You could also use this to ramp more or less aggressively.
More Cloud Regions
When planning your load test scenarios you might have noticed a special region called “Any Cloud Region” that showed up in the mid-Atlantic, and when you selected this region Loadster would automatically find a region with capacity for your test. Now, you’ve got your choice of three wildcard regions: “North America (Any)”, “Europe (Any)”, and “Global (Any)”.
By the way, we’ve also introduced new regions in Canada, Mexico, and New Zealand. This brings the total to 32.
Improvements & Bug Fixes
- Improved usability around disabling/enabling monitors.
- Improved usability around the script history tabs and splitter in the script editor.
- Fixed an occasional short-lived race condition with the Incidents tab’s open incident count.
- Fixed a bug that caused the top nav to show a running test after a failed launch.
- Fixed a bug with the Browser Console tab not showing up in Traces.
- Improved reliability of launching enormous tests when there are AWS capacity shortages.
- Improved UI consistency of the Timeline and modal screenshots.
2025.09
Playwright Test Support Beta
The Playwright testing feature is progressing nicely! This will enable the use of ordinary Playwright Test JS scripts for load testing and/or site monitoring, as an alternative or in addition to Browser Bots and Protocol Bots.
We look forward to a general release soon, but for now if you’d like to join the beta please contact help@loadster.com.
Improvements & Bug Fixes
- Added a section to the JSON exported test report with high-level metrics broken down by bot group.
- Fixed confusing date formatting on test reports with certain combinations of date preferences and timezones.
- Fixed a race condition that sometimes caused test reports to incorrectly show the test duration as 0:00:00.
- Improved multibyte character support in names of scripts, projects, etc.
Further Reading
Playwright Test support has been steadily progressing through experimental and early access phases, and this beta milestone brought it closer to general availability.
The JSON export improvements are particularly useful if you’re integrating Loadster into your CI/CD pipeline. With the Loadster CLI, you can trigger load tests from your build process and then parse the exported JSON report programmatically to check for regressions or threshold violations. Having bot group metrics in the export makes it easier to write automated checks that are specific to different parts of your scenario.
2025.08
Camera Device Simulation
If you’re testing applications that require access to a webcam device on the page (such as video conference or streaming platforms), Loadster now has a way for you to supply your own video files to simulate camera input. Please contact help@loadster.com if you’d like this enabled for your account.
Further Reading
Browser Bots drive real Chromium browser instances, which means they can interact with web
APIs that require hardware devices like cameras and microphones. Camera device simulation takes this a step further by
letting you supply custom video files that act as a virtual webcam feed. This is particularly useful for load testing
video conferencing platforms, telehealth applications, streaming services, or any web app that uses getUserMedia() to
access camera input.
When running load tests against these kinds of applications, each bot can present its own simulated camera feed, so the server-side experience closely mirrors what would happen with real users joining with their webcams on. This helps you catch performance issues related to media processing, encoding, and distribution under realistic conditions.
2025.07
Playwright Test Support Early Access
July has been dedicated to only one feature, which is Playwright Test support. This has been a considerable effort on our end, but should be very nice if you already have Playwright Test suites that you’d like to use for load testing and monitoring!
Please email help@loadster.com if you’d like to try it ahead of the public launch.
Bug Fixes
- Fixed a layout issue with the monitor status bars on a hosted status page, to make the bar segments more correctly proportioned.
- Improved time range precision on incident durations for human readability.
- Removed certain details from the logs that are sent out in monitoring pass/fail emails.
- Fixed a potential browser crash due to a memory leak related to date parsing!
- Corrected a discrepancy in overall load test network throughput measurements caused by the Chrome browsers themselves downloading updates at startup, unrelated to the actual traffic in the load test.
Further Reading
Playwright Test support gives you a different approach to scripting compared to Browser Bots. While Browser Bots let you record interactions visually and build scripts without writing code, Playwright Test scripts are written in JavaScript and give you full programmatic control. Both approaches use real Chromium browsers under the hood, so the load they generate is realistic either way.
2025.06
Features & Improvements
In June we added experimental support for running ordinary Playwright Test JS scripts in a load test, by request only for the time being! Please email help@loadster.com if you’d like it enabled for your account.
Other improvements:
- Dates and times are now displayed with clear timezone indication, showing the full timezone on hover in many places.
- The GCP cloud configuration now launches bigger instances by default, for better performance and throughput.
- The sidebar in the scenario editor is now sticky so it remains visible when scrolling scenarios with many bot groups.
Bug Fixes
- Fixed a problem with playing external scripts with the Loadster CLI.
- Fixed a bug in the onboarding overlay styles that affected new users.
- Reverted the shareable test report link to the old format.
Further Reading
The experimental Playwright Test support is a major change. Loadster already supports Protocol Bots for HTTP-level scripting and Browser Bots for recorded browser interactions, but Playwright Test scripts let you bring your own standard JavaScript test code and run it at scale. If your team already maintains Playwright Test suites for functional testing, you can now reuse those same scripts for load testing and site monitoring without rewriting anything.
The bigger default GCP instances in cloud regions mean more CPU and memory per engine, which translates to better performance and the ability to run more concurrent browser bots per instance. This is particularly relevant for browser-based and Playwright tests, which are more resource-intensive than protocol-level scripts.
2025.05
Improvements
- Made the monitoring location appear for each monitor cycle (on hover) when viewing historical monitoring results, so you can easily tell which location a cycle ran from. This can help in diagnosing regional or location-specific problems or differences in performance between regions.
- Added a breakdown of fuel usage by recent test to the Billing page, so you can see how much fuel each individual test consumed.
- Optimized query performance for several fuel-related queries, mainly affecting customers with lots of high-frequency monitors.
Bug Fixes
- Fixed a problem with graph legend column sorting being lost on refresh during a load test.
- Fixed an issue that sometimes caused a browser tab to crash when multiple users viewed the same test simultaneously when the final test report was being generated.
Further Reading
Loadster’s Site & API Monitoring feature runs your scripts on a schedule from multiple locations around the world. When something goes wrong, knowing which location flagged the issue is important – a failure from one region but not others can point to a regional CDN problem, DNS routing issue, or a data center outage rather than a global incident. The location-on-hover improvement makes that information immediately accessible without digging through logs.
The fuel usage breakdown on the Billing page gives you visibility into how your Loadster fuel is being consumed across individual tests. Fuel powers your load testing and monitoring, so understanding which tests use the most can help you plan and budget more effectively.
2025.04
Features & Improvements
- Introduced drag-and-drop to add steps to a script.
- Improved Loadster Engine stability in certain situations.
- Improved usability of the timezone setting.
- Clarified the use of phone numbers for monitoring alerts.
Bug Fixes
- Made the Protocol Bots GZIP response decoder less picky, to accommodate servers that truncate trailers.
- Fixed the syntax highlighter for variables and expressions to handle less common edge cases.
- Fixed a problem that caused double quote characters in expressions to be double-escaped in the editor.
Further Reading
Drag-and-drop for script steps means you can now arrange your browser scripts and protocol scripts more easily as you go, instead of adding new steps to the end and then reordering them.
Variables and expressions let you inject dynamic values into your scripts – things like timestamps, random numbers, or values pulled from datasets.
2025.03
Features & Improvements
- Introduced a more usable layout for the scenario editor.
- Added INFO traces for Protocol Bots.
- Added inline help to the script editor.
- Improved the usability of enabling and disabling monitors and adjusting frequency.
- Added cluster sizing to test reports.
Bug Fixes
- Fixed scrolling issues with the trace details modal dialog.
- Fixed incompatibilities in the Postman importer.
- Fixed a browser performance issue with the hosted status pages.
- Fixed a bug that caused browser instability when adding included scripts.
- Fixed a bug with duplicated projects losing references in scripts.
Further Reading
The scenario editor is where you configure how your load test actually runs – how many bots, which scripts they execute, and how traffic ramps up and down over time. The layout improvements in this release make it easier to see and adjust all of that at a glance, especially for scenarios with multiple bot groups.
The INFO traces for Protocol Bots give you a lightweight way to see what’s happening for a subset of bots when running a load test. They include full HTTP requests and responses, along with other details about the response.
2025.02
Features & Improvements
- Added a preview of the first lines in a large CSV dataset upload.
- Increased the Trace limits for smaller tests.
- Improved variable name validation hints.
- Improved scrolling and playback usability in the script editor.
- Added a prompt to append or replace when recording additional steps for an existing script.
Bug Fixes
- Fixed usability problems with copying and pasting steps between scripts.
- Fixed a bug with editing an HTTP request body.
- Fixed a UX problem with the team member invitation form when the seat limit was reached.
Further Reading
Datasets are how you feed dynamic data into your scripts – things like usernames, passwords, product IDs, or any other values that should vary between bots or iterations.
The increased trace limits for smaller tests mean you’ll get more detailed trace data when running tests that don’t generate massive volumes, which is helpful for debugging during the scripting and tuning phase. For more on using traces to understand what happened during a test, see analyzing test results.
2025.01
Improvements & Bug Fixes
- Fixed a bug that allowed users removed from the team to remain in monitor notification policies.
- Fixed a bug that caused the script playback panel to remain collapsed during playback in some situations.
- Improved the first-time scripting experience.
- Fixed a bug that caused “Wait For” step timing to show up in the test report.
- Added navigation guards to prevent accidentally losing unsaved changes to scripts.
- Improved bot browser stability and memory footprint.
Further Reading
The memory optimization work is about making each bot’s browser use less RAM, which directly affects how many bots can run on each clustered cloud instance when running load tests.
Browser memory bottlenecks can produce false negative results in which the test runs poorly and the site appears to be having trouble, when in fact some or all of the problems might stem from the browsers themselves being overburdened.
Loadster automatically sizes test clusters according to typical browser resource usage across the large majority of our customers’ sites, but there are times when you should select 2X or 4X cluster sizing to allocate additional resources.
2024.12
Updates from 2024.12
In December 2024, we polished the script editor UI, fixed graph legend visibility issues, and improved the Loadster Recorder browser extension.
2024.11
Updates from 2024.11
In November 2024, script recording was refined with better hover detection, the editor gained an Add button on hover, and Browser Bot timeouts improved.
2024.10
Updates from 2024.10
In October 2024, Loadster Monitoring Plans launched, the script recorder got a major update, and the browser script Timeline gained screenshot thumbnails.
2024.09
Updates from 2024.09
In September 2024, webhooks were extended to notify on load test start/finish with JSON reports, and the Recorder gained hover event recording support.
2024.08
Updates from 2024.08
In August 2024, monitors gained a quick enable/disable toggle, project search and sorting improved, and PDF invoices were added to billing emails.
2024.07
Updates from 2024.07
In July 2024, on-demand tracing via bot.trace() was added, weekly monitoring summary emails launched, and monitors gained misconfiguration alerts.
2024.06
Updates from 2024.06
In June 2024, free hosted status pages launched for monitors, wildcard cloud regions were introduced, and plan vs fuel prompts were added to test launch.
2024.05
Updates from 2024.05
In May 2024, Site & API Monitoring returned to Loadster, Slack integration gained per-channel notification control, and error counting was improved.
2024.04
Updates from 2024.04
In April 2024, Monthly Bot Plans were reintroduced with generous load testing allocations, and we optimized performance for large project lists.
2024.03
Updates from 2024.03
In March 2024, traces gained bot script logs for better debugging, billing was updated for 3D Secure payment flows, and cloud engine memory improved.
2024.02
Updates from 2024.02
In February 2024, Monthly Fuel Plans returned for longer-term testing needs, test reports gained step numbers, and percentile calculations were refined.
2024.01
Updates from 2024.01
In January 2024, the Project page was redesigned with tabbed views for scripts, datasets, scenarios, and test runs, with search and sort capabilities.
2023.12
Updates from 2023.12
In December 2023, the Loadster CLI 2.0 shipped with external script support, Browser Bots gained bandwidth throttling, and full page screenshots defaulted on.
2023.11
Updates from 2023.11
In November 2023, Browser Bot traces began including browser console logs, Chrome cache behavior improved, and several UI polish fixes were shipped.
2023.10
Updates from 2023.10
In October 2023, trace filtering gained multi-dimension support with exact and partial match modes, and Slack integration messages were improved.
2023.09
Updates from 2023.09
In September 2023, Slack integration started posting load test results, Browser Bots gained configurable geolocation, and JavaScript autocomplete was added.
2023.08
Updates from 2023.08
In August 2023, five new cloud regions were added across Australia, India, Indonesia, Spain, and Italy, and traces began including bot logs from the iteration.
2023.07
Updates from 2023.07
In July 2023, Browser Bots gained HTTP request blocking by URL pattern, and the script recorder improved selector generation for dynamic ID frameworks.
2023.06
Updates from 2023.06
In June 2023, manual transaction timers were added via bot.timer(), custom cluster sizing launched for CPU-heavy sites, and CSPs were disabled by default.
2023.05
Updates from 2023.05
In May 2023, Protocol and Browser Bots gained host-specific custom headers, the cloud region map added capacity indicators, and EmberJS selectors improved.
2023.04
Updates from 2023.04
In April 2023, we improved trace table pagination and filtering, separated trace details into multiple tabs, and improved large cluster launch reliability.
2023.03
Updates from 2023.03
In March 2023, scrollbar visibility improved, the script editor gained PageUp/PageDown support, and the trace table added pagination for large result sets.
2023.02
Updates from 2023.02
In February 2023, comment blocks were added to scripts for readability, the browser script memory leak was fixed, and the file upload limit was raised to 60MB.
2023.01
Updates from 2023.01
In January 2023, modifier key support was added to browser clicks and keyboard actions, enabling shortcuts like Ctrl+F in Browser Bot scripts.
2022.12
Updates from 2022.12
In December 2022, URL path simplification was added to test reports, the Next Steps section was improved, and projects gained a one-click duplicate feature.
2022.11
Updates from 2022.11
In November 2022, Browser Bots gained URL instrumentation for measuring background request performance, and trace limits were increased with column filtering.
2022.10
Updates from 2022.10
In October 2022, the scenario editor was redesigned, running bots gained a stacked area graph, new bot identification methods were added to code blocks.
2022.09
Updates from 2022.09
In September 2022, graph timestamps on hover were added, trace filtering gained search capability, and dedicated bare metal script players were deployed.
2022.08
Updates from 2022.08
In August 2022, bot groups gained delayed start for stair-step patterns, Browser Bots added keypress actions, and notification policies were improved.
2022.07
Updates from 2022.07
In July 2022, onboarding was simplified for new accounts, the script editor gained a side-by-side layout for Logs and Timeline, and login pages got faster.
2022.06
Updates from 2022.06
In June 2022, the test report gained a Request Count column, double-click was added to click steps, and HTTP operations began returning response objects.
2022.05
Updates from 2022.05
In May 2022, graph and table data became copyable for spreadsheet analysis, global custom headers were added, and the Vue 3 migration was completed.
2022.04
Updates from 2022.04
In April 2022, the project page gained a cloud region map and high-level test stats, and the Activity Feed was redesigned with more comprehensive test info.
2022.03
Updates from 2022.03
In March 2022, Browser Bots gained the ability to open new windows from code blocks, a silent step option was added, and report filtering improved.
2022.02
Updates from 2022.02
In February 2022, graph overlays were introduced to show metric relationships during and after tests, and auto-generated selector quality was improved.
2022.01
Updates from 2022.01
In January 2022, we extended the Traces TTL to 30 days, redesigned the Play buttons, improved script deletion messaging, and fixed several editor UI issues.
2021.12
Updates from 2021.12
In December 2021, single bot scripts began playing in AWS Lambda for reliability, self-service username changes were enabled, and timeout messages improved.
2021.11
Updates from 2021.11
In November 2021, Browser Bots gained camera, microphone, and file input support, and test reports added URL simplification options for cleaner metrics.
2021.10
Updates from 2021.10
In October 2021, clobber-resistant script editing launched with user badges and conflict detection, and Protocol Bots gained per-domain global headers.
2021.09
Updates from 2021.09
In September 2021, Browser Bots gained select-by-value/text/index and multi-window management, and Speedway launched for Site & API Monitoring.
2021.08
Updates from 2021.08
In August 2021, Browser Bot traces launched, capturing screenshots, resource waterfalls, and detailed error messages on bots with errors and first bot runs.
2021.07
Updates from 2021.07
In July 2021, Browser Bots gained Wait For, Select, and Hover step types, and scripts gained the ability to include other scripts for reusable flows.
2021.05
Updates from 2021.05
In May 2021, multiple team support launched, allowing user accounts to belong to more than one team with a team switcher in the dashboard.
2020.10
Updates from 2020.10
In October 2020, Browser Bots launched, controlling real headless Chromium browsers for realistic testing of complex website interactions and OAuth flows.
2020.07
Updates from 2020.07
In July 2020, new integrations launched for Slack, Opsgenie, and VictorOps, delivering quick monitor failure and recovery notifications to your team.
2020.06
Updates from 2020.06
In June 2020, the PagerDuty integration launched for Site & API Monitoring, enabling incident notifications, escalations, and on-call rotations.
2020.05
Updates from 2020.05
In May 2020, Loadster Workbench reached end of life. All efforts shifted to the web platform for a faster, more capable load testing experience.