TrueWatch released a new product update on August 19, 2026, with improvements across Toby AI TruePilot, Toby AI Agents, Agent Monitoring, Work Area, logs, Infrastructure, Dashboards, Generated Metrics, Application Performance Monitoring, Data Forwarding, integrations, DQL, and DataKit.
Toby AI TruePilot
Infrastructure analysis with page context
Context from infrastructure pages can now be passed to Toby AI TruePilot, and an AI Analysis shortcut has been added to those pages. An engineer looking at a host or container can ask for an analysis without restating what they are looking at.
Toby AI Agents
Self-observability settings take effect immediately
Changes to self-observability options now apply as soon as they are saved, with no need to restart or re-run the agent.
Agent Monitoring
New agent types recognized
Agent Monitoring now recognizes CodeBuddy, Spring AI Alibaba, DeepSeek Harness, and Cursor.
Work Area
Global Work Area for multiple open pages
A global Work Area has been added that keeps multiple feature pages open as tabs and switches between them without losing state.
- The current page can be added to the Work Area, and new pages can be opened through the module selector or the catalog.
- Up to eight work items can be open at the same time.
Logs
Log index cloning
A log can now be copied into a dedicated clone index, which gives one log stream a second retention policy without duplicating ingestion.
- After a log is written to the first standard index it matches, it can be copied asynchronously to a dedicated clone index based on additional conditions.
- When creating or editing a log index, the clone index name, retention period, and match conditions can all be configured.
- A clone index preserves the fields and tags of the original log in full. Each standard index can have at most one clone index.
- Once a clone rule takes effect it applies only to newly written logs and does not backfill historical data.
Wider search highlighting
Search highlighting is no longer limited to the message field. All content that matches the search condition is highlighted on the page. Analysis view switching preserves grouping After a BY grouping condition is set in the top list, switching to the list or stacked list view and back now preserves the previous grouping condition and the related analysis settings.
Advanced query compatibility
When an advanced query statement contains a BY condition, the page now ignores that condition to avoid conflicts with the analysis settings.
Infrastructure
Network device topology
A new network device topology diagram shows the physical connections between network devices.
Dashboards
Guaranteed streaming aggregation acceleration
A new Guaranteed Streaming Aggregation Acceleration setting has been added. When it is enabled, qualifying queries use streaming aggregation acceleration by default, and the system determines automatically whether a given query is eligible based on its characteristics.
Generated Metrics
Query acceleration list
A new query acceleration list shows the streaming aggregations that have guaranteed acceleration enabled, along with their source dashboard, chart, and related details.
Application Performance Monitoring
Default trace detail view
Preferences add a default trace detail view option, so trace details can open on either the waterfall chart or the flame graph.
Data Forwarding
Agent Monitoring as a data source
Data forwarding rules can now select Agent Monitoring as a data source.
TiDB Cloud Lake as a forwarding target
Observability data can now be forwarded to TiDB Cloud Lake in Parquet format.
Integrations
- TiDB Cloud Lake: a new integration that connects TiDB Cloud Lake directly as an external data source.
DQL
SHIFT time offset queries
Aggregate expressions now support <aggregate> SHIFT <duration> for referencing data from a historical window, which makes period-over-period comparison possible inside a single query.
Other
API Key notes in the list
The API Key list now displays note information alongside each key.
Trial workspace end-of-service notice
Users entering a feature module in a trial workspace now see an upgrade prompt noting that the trial edition stops service on September 2, 2026.
DataKit
New features
- The SNMP collector collects LLDP and CDP topology relationships, and adds a device_namespace tag and an interface_index field.
- The PostgreSQL collector supports checkpoint metrics on PostgreSQL 17 and later.
Improvements
- DCA 0.1.9 identifies the latest DataKit v1 and v2 versions separately.
- Improved compression, predicate digest inheritance, and large-packet splitting for tail-sampling data packets.
- Expanded documentation for Kubernetes node label global host tags and the log json_as_fields configuration.
FAQ
What is the biggest operational update in this release?
Log index cloning. A log stream can now be copied asynchronously into a dedicated clone index with its own retention period and match conditions, so a subset of logs can be kept longer, or isolated for a specific team, without duplicating ingestion. Clone rules apply only to newly written logs, so plan the rule before the data you want to keep arrives.
What changed for Toby AI TruePilot and Toby AI Agents?
Toby AI TruePilot can now receive the context of an infrastructure page and be launched from an AI Analysis shortcut on that page. For Toby AI Agents, changes to self-observability options take effect immediately instead of requiring a restart.
Why does this release matter for query performance?
Dashboards can enable guaranteed streaming aggregation acceleration, with the system deciding per query whether it applies, and the new query acceleration list under Generated Metrics shows which streaming aggregations are accelerated and where they came from. DQL also adds SHIFT, which moves period-over-period comparison into the query itself.