TrueWatch released a new product update on September 9, 2026, with improvements across Toby AI Agents, Agent Monitoring, Real User Monitoring, logs, Unified Catalog, Infrastructure, Dashboards, Pipeline, Event Center, Incident Center, Monitoring, Status Page, integrations, and DataKit.
Toby AI Agents
Enhanced task planning and automation
Plan Mode now produces clearer plans and execution steps. Automated tasks can be managed through natural language in task conversations, with form-based custom configuration for more efficient task creation and adjustment.
Agent Monitoring
End-to-end annotation and evaluation workflow
Annotation queues can be created from Trace query criteria to record scores, issue types, and reference answers, and to add samples to datasets. Datasets can be organized by Trace and manual annotation criteria, with batch evaluations that export to JSONL or CSV.
Clearer billing and quota rules
Agent Monitoring data in Commercial editions is billed by Trace or Span count. Enterprise editions include it in traffic and storage billing. Free editions apply usage quotas based on Trace count.
Improved user information display
The User dimension in Toby AI Agents can now be associated with workspace members and displays usernames, making users easier to identify and usage easier to analyze.
Real User Monitoring
Funnel drop-off analysis
Funnel Analysis now displays the number and rate of users lost between adjacent steps, with drill-down from lost Session details to Session Replay to help identify the causes of user drop-off.
Logs
Multi-index writes
The former clone index has been upgraded to independently managed multiple indexes, so the same log can be written to several indexes under different rules. An existing index configuration can be copied and new filters added. Each index can be configured, enabled or disabled, queried, and assigned a retention period independently. Existing clone indexes retain their identifiers, names, and data after the upgrade.
Unified Catalog
Centralized associated events
A new Events tab on the System Entity Details page aggregates events from constituent and dependent entities by default. Events can be filtered by entity type and name, with their sources displayed.
Infrastructure
Optimized network page hierarchy
The standalone Topology entry has been moved under Network Flow and is now displayed alongside List as a peer view. Existing functionality remains unchanged.
Dashboards
Bulk operations and parameterized sharing
Charts can be copied, moved into groups, or deleted in bulk. Share links now accept parameters that customize the default content displayed when opened.
Pipeline
Version management
Pipeline adds version management with version history, change comparison, rollback, and cloning of a selected version, giving teams more control over Pipeline changes.
Event Center
Event-to-incident visibility
Events associated with incidents now display incident indicators and relationship information, with quick filtering support.
Incident Center
Tencent Meeting war rooms
IM Apps now support Tencent Meeting integration. Once configured, a Tencent Meeting war room can be created directly from incident details to quickly start a meeting and coordinate incident response.
Monitoring
Easier mute rule filtering
Advanced filters for mute rules now include commonly used field options, helping users locate target rules more quickly.
Status Page
Broader service status coverage
Status monitoring now covers additional common modules, including Monitoring, Incident Center, and Error Center.
Other Improvements
Quick filters in viewers can now be refreshed manually to retrieve the latest available values. The update path for synthetic test tasks in environments with large volumes of historical data has been optimized, reducing update timeouts.
Integrations
New Apache Spark integration extends observability to big data processing scenarios.
Zero-code integration for LangChain Deep Agents adds OpenTelemetry-based best practices and setup documentation.
New Tencent Cloud CAT Synthetic Monitoring integration with collector, cloud collector template, dashboard, and monitor.
New Volcengine cloud collector templates: ALB, CLB, EIP, NAT, MySQL, PostgreSQL, Redis, MongoDB Replica Set, MongoDB Sharded Cluster, RabbitMQ, RocketMQ, VCM Kafka, NAS, TOS, VKE, ES Cloud, and SSL Certificate.
New Alibaba Cloud collector templates: ALB, AnalyticDB for MySQL, FC, MSE, VPC Peering Connection, and WAF.
New Huawei Cloud GeminiDB InfluxDB collector template.
New AWS collectors: Config, SES, and Amazon Bedrock.
Enhanced AWS SNS Collector with Topic properties, tags, and CloudWatch metrics.
Multi-cloud account balance monitoring for Alibaba Cloud, Volcengine, Tencent Cloud, and Huawei Cloud, with tiered balance alert monitors.
Updated integration documentation tags for easier discoverability.
Collector fixes
Fixed the Alibaba Cloud SAE Instance script ID and an Alibaba Cloud RDS collector parameter issue. Fixed incomplete data collection in the Kingsoft Cloud MySQL collector. Enhanced the Kingsoft Cloud KEC collector with batched collection and writes.
DataKit Updates
New features
Pipeline adds an optional Rust JIT execution engine for improved processing performance. DataKit Operator central election backend now supports completing elections within the Kubernetes cluster. DDTrace supports extracting numerical values from span.metrics via customer_tags. New sampling priority decision metrics have been added.
Bug fixes
Fixed an issue where SQL Server and Redis single-instance collection overwrote user-defined custom server tags.
Optimizations
Kubernetes resource metrics now use the informer cache. Browser-based synthetic testing upgraded Lightpanda to v0.4.0-g1. SNMP example configuration has been improved. Oracle collector validated for Oracle 26ai compatibility.
Bug Fixes
Fixed a trackInteractions / trackUserInteractions parameter mismatch during RUM Web application setup. Fixed inconsistent container counts between Host Details and the Container Explorer. Fixed an issue where Pipeline tests could not retrieve debug logs. Fixed an issue where the SQL Server database integration failed on its first run. Additional fixes for APM dashboards, Incident Center war room notifications, network log ingestion, and Pipeline name search.
FAQ
What are the biggest updates in this release?
The end-to-end annotation and evaluation workflow for Agent Monitoring is the standout addition. Teams can create annotation queues from Trace queries, score and tag each trace, build datasets, and run batch evaluations with JSONL or CSV export, closing the loop between agent operation and quality assessment.
What changed for log management?
Multi-index writes replace the former clone index with independently managed indexes. The same log can be routed to multiple indexes under different rules, each with its own retention period and filters, without mutual-exclusion restrictions on queries.
How does funnel drop-off analysis help with user experience?
Funnel Analysis now shows exactly how many users are lost between each step and at what rate. You can drill down from a lost session directly into Session Replay to see what went wrong, turning a high-level conversion metric into an actionable debugging workflow.