Reorganized the previously long, single-row menu into two levels: top-level categories (WAS, Database, Kubernetes, SLA, Alerts, Settings) → sub-categories (Dashboard, Applications, JVM, Transactions)
Even when you jump directly to a deep screen via a link, the left menu automatically expands to and highlights that location
Each category shows a badge with its instance count next to it, which stays visible even when the menu is collapsed
Automatic Detection of 16 T-Map Incident Patterns
Plots all transactions in a time window on a time (x-axis) × response time (y-axis) plane, and automatically detects the distinct shape each type of incident produces
The color of the status indicator circle shows severity — green (normal), yellow (minor), orange (warning), red (critical)
Even for the same error, guidance on where to investigate differs depending on response speed
The AI button next to the status circle shows, all at once: ① a one-line summary, ② the detected pattern, ③ likely causes, and ④ recommended menus to check
Location: WAS → T-Map
Automatic Diagnosis of 19 Transaction Root-Cause Patterns
Opening an individual transaction automatically diagnoses all 19 root-cause patterns at once, ranking them by severity color (red, orange, gray) at the top of the screen
Automatically classifies not only slow SQL, external calls, and errors, but also easy-to-miss patterns such as N+1 repeated queries, connection pool waits, and serialized external calls
Slow SQL and external calls are ranked together with elapsed time and their share of the total; clicking an item jumps to its position in the timeline below
Transaction details now open in a dialog (popup) instead of a separate page
The call flow is shown as a hierarchy diagram color-coded by type (database queries, external API calls, general methods, errors) alongside a timeline bar (waterfall), making it easy to spot slow segments
Location: WAS → Application Group / Instance / T-Map
Added an Overview tab to WAS > Application and JVM
Brings together 6–10 key metrics — throughput, response time, error rate, and more — on a single screen, so you can assess instance status in one click
Location: WAS → Application / JVM → the first "Overview" tab
Placed an AI help button in the same spot on every chart — metric tabs, chart headers, dashboard cards, and more — so you can check a metric's meaning and normal range right on the screen
The AI Performance Diagnosis button in the transaction dialog passes the same 19 automatic diagnosis results shown on screen directly to the AI, keeping what the operator sees consistent with the AI's response
Slow SQL Execution Plan Analysis and Index Recommendations
When slow SQL is detected, the AI automatically provides execution plan analysis and index recommendations
Analysis is tailored to the database type (MySQL, Oracle, and others)
Per-Account and Per-Permission-Group Control of CogentAI
CogentAI (the AI diagnosis feature) can now be turned on or off per account and per permission group, letting you enable AI features for specific groups only
Note: If CogentAI is disabled in a group's permissions, AI menus and buttons will not appear for users in that group. If any of your users were previously using AI features, check their group permissions.
Improved screen rendering and the responsiveness of chart and screen transitions
Switched server data exchange to a WebSocket-based approach
Lightened screen operation in environments with a large number of monitored agents
Consolidated repeated alarms into a single start notification, a single recovery notification, and periodic reminders, and grouped duplicate alarms from the same cause
Improved alarm processing performance and stability
Added traffic intensity shading to the transaction heatmap (T-Map), so that even within the same response-time band, periods with concentrated requests stand out by color depth, making it easy to spot traffic hotspots
Location: WAS → Dashboard / Application → T-Map
Applied by default with no extra setup required; the shading intensity can be set to strong (full saturation), medium (default), or subtle (muted)
Summarizes the status of multiple services on a single screen, letting you assess the operational state of all services at a glance
Displays call count, active users, error rate, TPS, average response time, and user satisfaction as time-series charts, along with changes compared to the previous day and the same day of the previous week
Changed T-Map error/warning cell borders to a bold solid line so they're clearly distinguished from normal cells, and adjusted the color of the top endpoint so it's no longer confused with genuinely critical ranges
Improved slow-query AI analysis to use tools matched to the database type, increasing analysis accuracy
Strengthened server cache management to improve stability during long-term operation
