> For the complete documentation index, see [llms.txt](https://docs.thousandeyes.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.thousandeyes.com/archived-release-notes/2026/2026-07-release-notes.md).

# Release Notes: July 2026

## 2026-07-30

### Endpoint Experience Events Are Now Generally Available

Endpoint Experience events are now generally available. Built on the ThousandEyes Event Platform, this feature helps you proactively detect employee-impacting issues across large Endpoint Agent deployments without creating and maintaining alert rules for every test or agent.

ThousandEyes automatically correlates Endpoint Agent test data to create organization-wide events that show impact, affected agents and tests, and severity. On the **Events** screen, you can review event summaries and details, then continue your investigation in the relevant Endpoint Agent view.

You can also create Event Detection alert rules and access Endpoint Experience events through the ThousandEyes API and ThousandEyes MCP server.

For more information, see [Event Detection](https://docs.thousandeyes.com/product-documentation/event-detection).

### Connected Devices Metrics in OpenTelemetry

Connected Devices metrics are now supported in ThousandEyes for OpenTelemetry. You can stream Connected Devices metrics to OpenTelemetry-compatible observability platforms alongside other ThousandEyes metrics, logs, and traces.

For a complete list of supported Connected Devices metrics and attributes in Data Model v2, see [ThousandEyes for OpenTelemetry Data Model v2 - Metrics](https://docs.thousandeyes.com/product-documentation/integration-guides/opentelemetry/data-model/data-model-v2/metrics).

### RoomOS Device Names Now Automatically Populate Endpoint Agent Names

The Endpoint Agent integration with Cisco RoomOS can now automatically set each agent name in the ThousandEyes platform to the RoomOS device name shown in Control Hub. When you enable **Obtain RoomOS name from ControlHub** in **Endpoint Experience > Agent Settings > Advanced Agent Settings**, the agent name becomes read-only in the ThousandEyes platform. To rename a RoomOS device while name syncing is enabled, update its name in Control Hub. ThousandEyes syncs the new name after the device's next check-in, making RoomOS devices easier to identify.

For more information, see [Naming RoomOS Agents](https://docs.thousandeyes.com/product-documentation/global-vantage-points/endpoint-agents/installing/cisco-devices/install-roomos#naming-roomos-agents).

## 2026-07-29

### Endpoint Agent Location Is Now Treated as Protected Data in Dashboards

Endpoint Agent location is now classified as personally identifiable information (PII) in dashboards, consistent with how it is handled across the ThousandEyes platform.

This affects existing dashboard embeds that include the **Map** or **Agent Status** widgets. Embeds are saved with **Anonymize PII data** enabled by default, and because location data is now anonymized, embeds saved with this default setting will no longer display data in these two widgets.

Viewing these widgets directly in a dashboard is not affected. This change applies only to embeds, based on the **Anonymize PII data** option selected when the embed was created.

To restore data in an affected embed, recreate or edit the embed with **Anonymize PII data** unchecked. This requires the Endpoint PII permission.

### Agent ASN and Geolocation Live Updates

Previously, Enterprise Agents only determined their ASN and geolocation during initial registration, based on their public IP address. If the agent's public IP address changed after installation, the reported ASN and geolocation remained unchanged. With this update, ThousandEyes now automatically refreshes an agent's ASN and geolocation based on its public IP address. This ensures that the agent's reported network provider and location remain accurate over time, even if they are moved or their ISP reassigns their IP address, improving inventory accuracy and making network troubleshooting more reliable without requiring manual intervention.

The system intelligently manages refresh timing to prevent "noise" from temporary network shifts and maintain stability. Once a change is confirmed, the new data will appear in your dashboards and OpenTelemetry feeds within 30 minutes, and in Event Detection within 24 hours. Geolocation updates will align with the standard location list.

## 2026-07-28

### HTTP Server and API DNS Lookup Duration for OpenTelemetry Data Model v2

ThousandEyes for OpenTelemetry now streams the `dns.lookup.duration` metric for HTTP server test results and API test step results in Data Model v2 when DNS timing data is available. The metric includes the `dns.question.name` attribute, which identifies the DNS hostname that the test resolves. Use this metric to monitor how long HTTP server tests spend resolving the target hostname and API test steps spend resolving the request hostname, alongside availability, throughput, response time, API step duration, and completion metrics.

For the full list of metrics and attributes, see [Network & App Synthetics Tests](https://docs.thousandeyes.com/product-documentation/integration-guides/opentelemetry/data-model/data-model-v2/metrics#network-and-app-synthetics-tests).

## 2026-07-27

### ThousandEyes MCP Server for ThousandEyes for Government

The ThousandEyes MCP server is now available in the ThousandEyes for Government instance. The MCP server lets you connect a compatible AI assistant to the ThousandEyes platform and use natural language to access supported monitoring data through a standardized interface.

{% hint style="info" %}
The ThousandEyes MCP server is not a generative AI (GenAI) feature.
{% endhint %}

Some MCP server tools are not available in the ThousandEyes for Government instance. For setup instructions and the full list of MCP server tools, see [ThousandEyes MCP Server](https://docs.thousandeyes.com/product-documentation/integration-guides/thousandeyes-mcp-server). For Government-specific availability and limitations, see [ThousandEyes for Government](https://docs.thousandeyes.com/product-documentation/thousandeyes-for-government).

## 2026-07-24

### Support for ThousandEyes Enterprise Agents on Raspberry Pi 5

ThousandEyes Enterprise Agents can now be installed on Raspberry Pi 5. In addition, the Raspberry Pi Enterprise Agent image has been upgraded to Ubuntu 24.04 LTS, providing enhanced security and long-term stability.

**Known Issues**

* Platform Display: The **Agent Settings** page currently lists "Raspberry Pi 4" as the only available platform option. This is a cosmetic labeling issue; the provided image is fully compatible with and supported on Raspberry Pi 5 hardware. A UI update is planned for an upcoming release.
* IPv6 Auto-Configuration: In IPv6-only environments using automatic address assignment, only SLAAC with static DNS configuration is currently supported. Support for automatic address assignment via DHCPv6 is planned for a future update.

## 2026-07-23

### Cisco Application Hosting Framework (CAF) version 5.2.0

CAF version 5.2.0 has now been released. This release is intended to replace the 5.1.3 image on all supported Cisco devices, and migrates the underlying operating system to Alpine Linux 3.23.

For upgrade instructions, see [Upgrading Enterprise Agents](https://docs.thousandeyes.com/product-documentation/global-vantage-points/enterprise-agents/managing/upgrading-enterprise-agents#cisco-application-hosting).

**New Features**

* Post-quantum cryptography (PQC) support.
* A new run-opt option to disable BBOT runs by default.

**MD5 Checksum Verification**

ThousandEyes recommends validating the checksum of downloaded packages. The md5 checksums for the CAF version 5.2.0 images are:

* **thousandeyes-enterprise-agent-x86\_64-5.2.0.cisco.tar**: 4bdb531967105baf79ee8d7f9244eeb1
* **thousandeyes-enterprise-agent-aarch64-5.2.0.cisco.tar**: 94e143355371ec7b6df741e467615b02

## 2026-07-16

### Chromium 150 Available for Browser Test Validation

Chromium 150 is now available for validation through the Dual Chromium option. For page load and transaction tests, set **Chromium Version** to **Newer when available** to test important scripts against the newer browser track before it becomes the default. The current version of Chromium remains Chromium 148 for **Default**.

Validate business-critical scripts before August 12, 2026, and review new errors or script behavior changes on the newer track before the **Default** version of Chromium is upgraded shortly after the validation period.

ThousandEyes also runs automatic validation before the **Default** Chromium upgrade for tests that can use Dual Chromium. These instant tests might appear in your **Activity Log**, but they do not consume units. For details, see [Configuring Dual Chromium](https://docs.thousandeyes.com/product-documentation/browser-synthetics/dual-chromium-option/configure-dual-chromium#thousandeyes-validation-on-dual-chromium-agents).

For Enterprise Agents, the agent must use the **Dual Chromium Version** BrowserBot package to run tests on the newer Chromium track. All Cloud Agents can run the newer track without additional agent setup.

For more information, see [Dual Chromium Option](https://docs.thousandeyes.com/product-documentation/browser-synthetics/dual-chromium-option).

### BrowserBot Test Containers Use User Namespaces on Native Podman Deployments

BrowserBot release 3.29.0 enables user namespaces for BrowserBot test containers that run with the Podman runtime on native Enterprise Agent deployments. This change does not apply to CAF or Docker deployments. User namespaces map container root to a non-root user on the host, which improves container isolation. No action is required.

On user-namespace-enabled native deployments, BrowserBot rootfs directories under **/var/lib/te-browserbot/te-browserbot-test-container/rootfs**\* might show ownership as **50000000:50000000** instead of **0:0**. This ownership is expected. Docker-based deployments continue to show **0:0**.

### UPDATE to Firewall Update Required for ThousandEyes Virtual Appliance (TEVA) and ThousandEyes Physical Appliance (TEPA) - Delayed Release of ThousandEyes Docker Adapter

In the June 1st, 2026 changelog ([Firewall Update Required for ThousandEyes Virtual Appliance (TEVA) and ThousandEyes Physical Appliance (TEPA)](https://docs.thousandeyes.com/release-notes/2026/2026-06-release-notes#id-2026-06-01)), we announced that we will be adding the **docker-ce** package to our ThousandEyes Virtual Appliance (TEVA) and ThousandEyes Physical Appliance (TEPA) images in preparation for the BrowserBot Docker adapter release planned for mid-July. This change is part of our continued efforts to improve the architecture of our BrowserBot tests, specifically as part of our transition from using the Podman adapter to a Docker-based adapter.

At this time, we have decided to delay the rollout of both the **docker-ce** package and the Docker-based adapter. You can either maintain your current firewall configuration as is (with the addition of `download.docker.com`), or roll it back, depending on your organization's security policies.

When we are ready to roll out the **docker-ce** package and the Docker-based adapter, we will provide at least 30 days notice before you need to update your firewall configuration again with `download.docker.com`.

If you have any questions, please contact ThousandEyes Support.

### Cut Down on Alert Noise with Alert Optimizations (Public Beta)

Alert Optimizations is now available as a public beta. Alert Optimizations helps teams reduce alert noise without manually reviewing every alert rule. ThousandEyes reviews recent alert activity, finds Network & App Synthetics manual alert rules that trigger and clear repeatedly in short cycles, and recommends switching those rules to Adaptive Alerting. In the ThousandEyes platform, the changes recommended by Alert Optimizations can reduce alert noise by up to 80%.

You can use these recommendations to focus operations teams on alerts that are more likely to need action. Recommendations appear in **Alerts > Active Alerts**, **Alerts > Alert History**, and **Alerts > Alert Rules**. On **Alerts > Alert Rules**, use the **With Optimizations** filter to review flagged rules. Open a rule to review the recommendation, then select **Apply Changes** to switch the rule to Adaptive Alerting for all associated tests, or select **Reject** to decline it.

For more information, see [Alert Optimizations](https://docs.thousandeyes.com/product-documentation/alerts/alert-optimizations).

## 2026-07-14

### Enhanced Provider Visibility in Provider Intelligence

You can now view providers with "no coverage" in the **Summary View** of Provider Intelligence results. By default, we show you a comprehensive list of all providers present in your selected locations, even those where performance data is currently unavailable.

This update increases transparency, allowing you to quickly identify the full scope of available providers at a glance, regardless of query parameters.

For more information, see [Data Sources and Coverage](https://docs.thousandeyes.com/product-documentation/internet-insights/provider-intelligence/data-sources-and-coverage).

## 2026-07-13

### Endpoint Agent Client Version 2.51.1

The following enhancements have been made to the Endpoint Agent client:

#### Windows and macOS

* When testing the underlay of an AnyConnect (ASA) tunnel, the Endpoint Agent no longer uses TCP SACK. This change reduces the risk of dropped user packets in deployments where many clients connect to a single ASA.

## 2026-07-09

### Heatmap Dashboard Widget

You can now add a **Heatmap** widget to custom dashboards. The widget shows a color-coded matrix where each cell represents one combination of a row dimension and a column dimension, such as **Test** by **Agent**. Cell color shows how the value compares with other cells in the same widget for the dashboard time range, helping you scan for hotspots and outliers across your deployment without opening each test or agent separately.

![Heatmap dashboard widget example](https://1112912342-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-M4QARF6s57qxMrOHDTZ%2Fuploads%2Fgit-blob-51917d50ac5c7faa72a6a8cd70c6a7080e323723%2Fdashboard-heat-map-widget-example.png?alt=media)

Use a heatmap when you need to compare two dimensions at once. A single heatmap can replace a large grid of separate **Number** or **Color Grid** widgets for questions such as which agents are slow on which tests, or how interfaces compare across devices.

For more information, see [Dashboard Widgets](https://docs.thousandeyes.com/product-documentation/dashboards/dashboard-widgets#heatmap-widget).

### Pie Chart and Stacked Area: Stack by Submetrics or by Group

**Pie** and **Stacked Area** widgets now support metrics from all data sources in the metric picker.

For composite metrics, such as **HTTP Response Time**, the default is **Stack By Submetric**. In widget settings, you can switch to **Single Stacked Chart** to combine submetric components and divide the chart by your **Group By** dimension, such as **Test**, **Agent**, or **VPC**.

For non-composite metrics, the chart divides by your **Group By** dimension.

For more information, see [Dashboard Widgets](https://docs.thousandeyes.com/product-documentation/dashboards/dashboard-widgets).

## 2026-07-07

### Endpoint Local Network Wireless Metrics for OpenTelemetry Data Model v2

ThousandEyes for OpenTelemetry now streams wireless, connection, and signal metrics for Endpoint Experience local network data using Data Model v2. You can monitor active Wi-Fi, wired, and cellular connections with metrics such as signal quality, throughput, link speed, connection score, and cellular signal strength indicators.

The `thousandeyes.source.agent.connection.type` attribute is deprecated. Use `network.connection.type` for all Endpoint Local Network metrics.

For the full list of metrics and attributes, see [Endpoint Experience Local Network](https://docs.thousandeyes.com/product-documentation/integration-guides/opentelemetry/data-model/data-model-v2/metrics#endpoint-experience-local-network).

### Endpoint Local Network System Metrics for OpenTelemetry Data Model v2

ThousandEyes for OpenTelemetry now streams additional Endpoint Experience local network system metrics using Data Model v2. The new metrics are `system.battery.charge`, `system.battery.health`, and `system.disk.free.utilization`, expanding system visibility beyond CPU, memory, and agent score.

The `thousandeyes.source.agent.serial_number` attribute is also reported on Endpoint Local Network metrics when available, helping downstream systems correlate the telemetry with the source device.

For the full list of metrics and attributes, see [Endpoint Experience Local Network](https://docs.thousandeyes.com/product-documentation/integration-guides/opentelemetry/data-model/data-model-v2/metrics#endpoint-experience-local-network).

## 2026-07-06

### Mobile Endpoint Agent Android Version 1.8.0

The Mobile Endpoint Agent includes the following enhancements:

* Updated the open source license declarations to reflect the latest libraries used by the agent.
* Added fallback scheduling to improve reliability when exact alarms are unavailable.
* Fixed an issue where HTTP connection timeouts could be incorrectly reported as HTTP 400 errors.
* Updated registration error messages to be clearer when network connectivity is unavailable.

## 2026-07-02

### Endpoint Agent Client Version 2.51.0

The following enhancements have been made to the Endpoint Agent client:

#### All Platforms

* Improved path trace algorithms to increase the success rate for traces that reach the target.
* Fixed an issue where the Endpoint Agent could fail to report the Zscaler VPN profile for endpoints using newer Zscaler clients. The agent now recognizes updated Zscaler gateway log entries, so Zscaler-based dynamic tags and path visualization reflect the connection as expected.
* Fixed an issue that could allow agents on the **NONE** release track to update unexpectedly.

#### Windows

* Fixed a crash caused by an invalid proxy bypass list.
* Fixed an issue that could cause aggressive process relaunches and high CPU usage if the **te-user-agent** process fails to launch.

#### macOS

* Enhanced the Endpoint Agent to support the latest system memory reporting in the upcoming macOS 27 release.

## 2026-07-01

### ThousandEyes MCP Server Works With GenAI-Powered Capabilities Disabled

We've removed agentic capabilities from the ThousandEyes MCP server, so it now functions purely as a standardized, permissioned interface to your ThousandEyes data. The server runs no Cisco or ThousandEyes generative models, does not send your data to a Cisco or ThousandEyes LLM, and does not use your data to train or fine-tune any model.

Because the server no longer performs any generative AI (GenAI), you no longer need to opt-in to ThousandEyes GenAI features to use it. Organizations that have GenAI features disabled, including those that have opted out of GenAI training on their data, can now connect an AI client to the MCP server while keeping those settings in place. Access is governed by the same controls as the ThousandEyes API: API Access permission, API tokens, or the OAuth 2.0 flow, and rate limits.

This gives teams in regulated environments a way to bring AI-assisted network intelligence into their own approved, auditable toolchain, under their own AI governance, without opting into any Cisco-operated generative AI.

{% hint style="info" %}
This reflects the MCP server's capabilities today. Organizations that have opted into GenAI data usage may, in the future, see agentic capabilities surfaced through the MCP server. Organizations that remain opted out will continue to access it as the standardized, non-generative data interface described here.
{% endhint %}

No action is required. For more information see the [ThousandEyes MCP Server](https://docs.thousandeyes.com/product-documentation/integration-guides/thousandeyes-mcp-server) documentation.

### DNS Response Details Now Available for HTTP Server Results

HTTP server test results can now show DNS response details when DNS response data is available. These details provide resolver-level evidence from the agent's DNS phase, including the resolver used, protocol, response status and return code, timing, truncation state, and returned records.

Use these details to distinguish DNS resolution issues from later HTTP, SSL/TLS, network, or application failures. For DNS-phase failures, the **Error Details** side panel can show **DNS Responses**. For successful HTTP server results, the **Response Details** side panel can show a **DNS Response** tab when DNS response data is present.

For more information, see [HTTP Server Tests](https://docs.thousandeyes.com/product-documentation/tests/http-server-tests) and [Using the HTTP Server View](https://docs.thousandeyes.com/product-documentation/internet-and-wan-monitoring/viewing-data/using-the-http-server-view).

### Agent API Serial Number Reporting for Cisco Devices

The agent API now includes a serial number field for Cisco devices hosting ThousandEyes Enterprise Agents. This gives API consumers a stable hardware identifier for Enterprise Agents that run on supported Cisco platforms, helping teams correlate agent inventory with device inventory and troubleshoot from API data with less manual lookups.

For more information about agent API changes, see the [ThousandEyes API changelog](https://developer.cisco.com/docs/thousandeyes/api-changelog/).

### ThousandEyes Support & Lifecycle Updates: July 2026

#### Recorder IDE End of Life Reminder

[As previously announced](https://docs.thousandeyes.com/release-notes/2026/2026-05-release-notes#id-2026-05-19), the Recorder IDE v1.24.0 will reach its Functional End of Life (EOL) date on July 26, 2026.

As of this date, the Recorder IDE will enter an unsupported and unmaintained state. While the Recorder IDE may continue to operate beyond this date, ThousandEyes makes no warranties or guarantees regarding its functionality, compatibility, or security. Customers are strongly encouraged to evaluate and transition to alternative solutions prior to this date.

#### Rocky Linux 9.7 End of Life Reminder

Rocky Linux 9.7 will reach its End of Life date on July 29, 2026. Enterprise Agents must be upgraded to a supported operating system version before then to remain functional.

#### Alpine Linux 3.21 End of Installation Support Reminder

Alpine Linux 3.21 will reach End of Installation Support on August 1, 2026. After this date, ThousandEyes will no longer permit new Enterprise Agent installations on this version of the operating system. Existing agents will continue to function and receive support until the End of Support phase begins.

{% hint style="info" %}
This applies to ThousandEyes Enterprise Agents deployed as Docker containers or using the Cisco Application Hosting Framework (CAF).

ThousandEyes does not support the native installation of ThousandEyes Enterprise Agents on Alpine Linux.
{% endhint %}

For upgrade instructions, see [Upgrading Enterprise Agents](https://docs.thousandeyes.com/product-documentation/global-vantage-points/enterprise-agents/managing/upgrading-enterprise-agents).

#### Amazon Linux 2 End of Support Reminder

[As previously announced](https://docs.thousandeyes.com/release-notes/2026/2026-04-release-notes#id-2026-04-10), Amazon Linux 2 entered the End of Support phase on June 30, 2026. ThousandEyes no longer provides support or guarantees software updates, including bug fixes, for Linux package agents running on Amazon Linux 2.

The End of Life date is August 30, 2026. ThousandEyes Enterprise Agents on Amazon Linux 2 must upgrade to Amazon Linux 2023 or another supported OS before then to remain functional. For upgrade instructions, see the [AWS documentation](https://docs.aws.amazon.com/linux/al2/ug/prepare-for-al2023.html).

### New Cloud Agents

New Cloud Agents have been added in the locations listed below. For a full list of Cloud Agents, see [ThousandEyes Cloud Agent Locations](https://www.thousandeyes.com/product/cloud-agents).

#### Alibaba

* Johor, Malaysia (Alibaba ap-southeast-8)
* Mexico (Alibaba na-south-1)
* Paris, France (Alibaba eu-west-2)

#### AWS Regions

* New Zealand (AWS ap-southeast-6)
* Taipei, Taiwan (AWS ap-east-2)

#### AWS Local Zones

* Honolulu, HI (AWS Local Zone) \[us-west-2-hnl-1a]
* Istanbul, Turkey (AWS Local Zone) \[eu-central-1-ist-1a]

#### AWS Wavelength Zones

* Casablanca, Morocco (Orange - AWS Wavelength) \[eu-west-3-cmn-wlz-1a]
* Dakar, Senegal (Orange - AWS Wavelength) \[eu-west-3-dss-wlz-1a]

#### Azure

* Brussels, Belgium (Azure belgiumcentral)
* Copenhagen, Denmark (Azure denmarkeast)
* Jakarta, Indonesia (Azure indonesiacentral)
* Kuala Lumpur, Malaysia (Azure malaysiawest)
* Santiago, Chile (Azure chilecentral)
* Vienna, Austria (Azure austriaeast)

#### GCP

* Bangkok, Thailand (GCP asia-southeast3)
* Querétaro, Mexico (GCP northamerica-south1)

### ThousandEyes Developer Hub Launches on Cisco DevNet

The new [ThousandEyes Developer Hub](https://developer.cisco.com/thousandeyes/) is now live on Cisco DevNet, giving developers and operations teams a dedicated starting point for building with ThousandEyes APIs, automation tools, telemetry integrations, and workflow examples.

Use the hub to find developer resources such as the API v7 developer reference, getting-started guides, automation tooling, and integration guidance from a single DevNet entry point.

For more information, see [ThousandEyes Developer Hub](https://developer.cisco.com/thousandeyes/).

### Bulk Editing in Test Settings

You can now update the test interval and assigned agents for multiple Network & App Synthetics tests from **Network & App Synthetics > Test Settings**. Select the tests you want to change, then use **Edit > Test Interval** to change how often eligible tests run, or **Edit > Manage Agents** to add or remove agents.

Bulk updates are submitted as jobs, so admins can update large sets of tests without opening each test individually. Before you submit a bulk interval change, the workflow flags tests that do not support the selected interval. Before you submit agent changes, you can review which agents will be added to or removed from each test.

For more information, see [Working with Test Settings](https://docs.thousandeyes.com/product-documentation/tests/working-with-test-settings#updating-multiple-tests).

### Cloud Insights Updates

#### AI Assistant Support for Cloud Insights Events

The Cisco AI Assistant now supports AI-assisted analysis for Cloud Insights inventory events and cloud traffic events (currently in Beta - see [previous announcement](https://docs.thousandeyes.com/release-notes/2026/2026-03-release-notes#id-2026-03-27)). When events are visible in the **Events** table, select **Summarize Events** to generate a consolidated analysis of all visible events, including an overview, key timestamps, possible causes or configuration changes, potential impacts, and recommendations.

To summarize both inventory events and cloud traffic events, go to **Cloud Insights > Views**. For individual inventory event analysis, go to **Cloud Insights > Inventory**.

After the initial summary, you can ask follow-up questions in the AI panel. For more information, including limitations such as the number of events the summary can handle, see [Events in Cloud Insights](https://docs.thousandeyes.com/product-documentation/cloud-insights/views/events-in-cloud-insights#ai-assisted-event-analysis).
