# New Features in 1.17 — 2026 Autumn Release

## Intelligence Chat Assistant

– **AI Chat — Floating chat assistant panel** — A floating chat widget is available to all logged-in users via a FAB button in the bottom-right corner. Ask natural-language questions about your PASOE environment and get instant answers: worst health probes, high HTTP thread saturation, high memory usage, session counts, recent alerts, PASOE status, and environment summaries.

## Rules Engine based AI Reports

– **AI Reports — Scheduled report generation** — A new Reports page allows admin/super users to create AI-powered reports on configurable schedules (once, interval, cron). Each report stores its full result and run history.

– **AI Reports — Email delivery** — Reports can be sent by email via configured SMTP. The email body contains a colour-coded health badge and recommendations; the full report is attached as an HTML file.

– **AI Reports — Slack delivery** — Reports with a health assessment or recommendations are delivered to a Slack channel using Block Kit: header, health badge, recommendations section, and context footer. No message is sent when there is nothing actionable to report.

– **AI Reports — Teams delivery** — Reports are delivered to Microsoft Teams via webhook using MessageCard format: coloured theme, health facts row, and recommendations bullet list.

– **AI Reports — Run History tab** — A dedicated tab shows all report execution records across all reports (admin/super) or own reports (regular users) in a sortable, filterable Clarity datagrid.

– **AI Reports — Timeline with inline previews** — The Reports Timeline tab shows runs with inline expand, result first-line preview, relative timestamps, left-border colour coding, and expand/collapse toolbar.

– **AI Reports — Access controls** — Email/Slack/Teams delivery configuration is restricted to admin/super users. “Restart All” and “Stop All” scheduler buttons are restricted to admin/super; regular users use per-row start/stop actions.

## Proactive Dashboard

– **Proactive Dashboard — New page** — A new `/proactive` page (bolt icon in the nav bar) shows AI report results as auto-refreshing cards. Each card shows report name, schedule badge, health status, last-run timestamp, result summary, and a detail modal (eye icon). Card border colour reflects the health assessment.

– **Proactive Dashboard — Shared reports** — Admin/super users can share reports with specific users or all users. Shared reports appear on the recipient’s Proactive page.

– **Proactive Dashboard — Health assessment** — Each report result carries a `result_health` value (ok / marginal / critical) reflecting the monitored system’s health, independent of execution status. Actionable recommendations are shown in the detail modal.

– **Proactive Dashboard — Dual status badges** — Cards show both the AI health assessment badge and the execution result badge, each with a tooltip explaining the distinction.

– **Proactive Dashboard — “Shared” filter toggle** — Toggle to show or hide reports shared by others.

## Integrations

– **Settings — SMTP email configuration** — A new SMTP tab (admin/super only) configures the outgoing mail server: host, port, user, password (stored encrypted, shown as set after save), TLS mode, and sender address. An inline “Test” button sends a test email and displays the result.

– **Settings — Slack & Teams integrations** — A new Integrations tab configures webhook-based Slack and Microsoft Teams integrations. Supports a default webhook URL with optional per-type overrides (reports / alerts / events) and per-type send toggles.

– **Settings — AI Providers tab** — Configures the AI provider for chat and reports. The built-in Rules engine is the active provider for 1.17; external LLM provider support (Anthropic, OpenAI, Ollama) is prepared for future releases.

## Intelligence Insights

– **Intelligence — Predictive Recycle Insights tab** — A new tab shows bar charts of predictive recycle events over time per PASOE, complementing the existing Anomaly Detection Insights tab.

– **Intelligence — Configurable critical window** — The `predictive_recycle_critical_minutes` threshold (default: 10 minutes) is now configurable system-wide in Settings → Intelligence and can be overridden per PASOE in Setup.

– **Intelligence — Auto-recycle on threshold** — PASOEs can be configured in Setup to automatically trigger a recycle when the predictive recycle job detects threshold breach, with an optional “Only recycle agents that have individually exceeded their threshold” flag.

– **Intelligence — Timeline UX overhaul** — Both Anomaly Detection Timeline and Predictive Recycle Timeline tabs have been rewritten with inline expand, relative timestamps, left-border colour coding, severity icons, and expand/collapse toolbars.

## Alerts

– **Alerts — Comprehensive audit grid** — The Audit component gains a full Alerts tab with server-side filtering (Env→Customer→PASOE cascade, time window, type, critical-only, adjustable max rows including Unlimited), “Delete selected”, and maintenance controls (purge older than, delete all).

– **Alerts — Server-side search** — A new `/alerts/search` endpoint supports server-side filter, sort, and configurable row cap (500–Unlimited).

– **Alerts — Inline health & threshold context** — The Feed/Events Alerts timeline shows inline `Health: X.XX` for Health Scanner alerts and `~X min to threshold` / `Threshold exceeded` for Predictive Recycle alerts.

– **Alerts — View details modal** — Both the Audit grid and Health page Alerts section support a “View details” action that fetches and displays the full alert record in a modal.

## Events

– **Events — Timeline UX overhaul** — Alerts and Actions timelines have been rewritten with inline expand, relative timestamps, colour-coded left borders, severity/action icons, and per-tab expand/collapse toolbars.

– **Events — Critical and Errors toggles** — Alerts timeline has an “Only critical” toggle; Actions timeline has an “Errors only” toggle.

– **Events — Stream limit with Max selector** — A configurable Max rows selector (100–Unlimited) with a cap warning when the limit is reached.

– **Events — Cascading Env filter** — An Environment selector has been added to the filter bar, cascading to Customer.

– **Events — Filter persistence** — Filter bar state is persisted to localStorage and restored on return.

– **Events — Auto-apply filters** — Filters apply automatically on change (debounced); the Apply button has been removed.

## Scheduler

– **Scheduler — Agent Maintenance Insights filter bar** — A unified filter bar for both Recycle and Cleanup Insights with Env/Customer/PASOE/ABL App cascade dropdowns applied server-side, a reload icon, reset icon, and inline Help panel.

## Dashboard

– **Dashboard — PASOE-level memory sparklines** — Per-PASOE memory sparklines (sum of all ABL app agents) with colour-coded trend indicator and click-to-open trend modal.

– **Dashboard — Contextual help icons** — Large-mode cards show `?` help icons (Clarity tooltip) for Host Summary, Health bars, and ABL Session Metrics sections.

## Health

– **Health — PASOE reachability status indicators** — The PASOE Instances grid on the Host view shows a tristate circle dot (grey=unknown, green=reachable, red=unreachable), proactively checked on page load. Dynamic PASOE tabs show a “PASOE unreachable.” message when the probe fails.

– **Health — Configurable alerts time window** — The Alerts section on the PASOE view now has a Show time window selector (All time / 1h / 4h / 12h / 24h / 7d / 30d / 90d).

## Backend

– **PASOE-level predictive recycle** — The predictive recycle scheduler and AI rules engine now track memory at PASOE total level (sum of all ABL apps). Alerts, recommendations, and auto-recycle logic all operate on the PASOE memory budget.

– **Gated Slack/Teams delivery** — Slack and Teams report notifications are only sent when the report has a health assessment or recommendations; no notification noise for reports with no actionable signal.

## Infrastructure

– **Database — Migration scripts** — A `migrations/1.17/` folder of idempotent per-table SQL scripts with `CREATE TABLE IF NOT EXISTS` / `ADD COLUMN IF NOT EXISTS` guards enables safe upgrades from 1.16 installations.

– **Compose Stack — Maintenance scripts** — A full set of ProDiscoverer Docker compose stack maintenance scripts added.

—

# New Features in 1.16 — 2026 Summer Release

## Frontend

– **Dashboard — Health Scanner bars large-mode only** — Health Scanner probe bars are now hidden in slim card mode and only visible when cards are expanded to large mode, reducing clutter in the default view.

– **Dashboard — Conditional chart buttons** — Memory vs HTTP Requests and ABL Sessions vs HTTP Requests chart buttons are now only shown when at least one of the relevant probe sources (Health Scanner or OEManager) is enabled for the PASOE.

– **Dashboard — Anomaly Detection graph improvements** — The Anomaly Detection graph now defaults to a 1-hour window; the time range dropdown matches all other charts (1–24 hours). A “No anomalies detected” message is shown when there is no data in the selected window.

– **Dashboard — Ping status dot repositioned** — The PASOE ping status dot has been moved into the PASOE name row (no longer occupies a separate card row). It is only shown when Health Check is enabled and a Health Check URL is configured for the PASOE.

– **Health — Agent Statistics reset modal** — Resetting agent statistics now shows a success or error info modal with the agent ID and PID, replacing the silent no-feedback behaviour.

– **Health — Larger operation modals** — Recycle Agents, Application Properties, and Session Manager Metrics result modals are now opened at large (`lg`) size for better readability of longer messages.

– **Service Cleanup — “All” option** — A new **All** period option has been added to each cleanup row. Selecting it uses `TRUNCATE TABLE` instead of a row-by-row DELETE, providing instant deletion of all records regardless of table size.

– **Service Cleanup — Batch DELETE** — Time-range deletions now use a batched approach (5 000 rows per batch with a timestamp-window strategy) instead of a single large DELETE, preventing database timeouts on tables with millions of records.

– **Header — Changelog** — The “What’s New” menu item has been replaced with “Changelog”, which opens a modal showing the full accumulated version history across all releases.

—

# New Features in 1.15 — 2026 Spring Release

## Frontend

– **Service page** *(super users only)* — new navigation item visible only to super-user accounts.

– **Service Cleanup tab** — delete records older than a selected period (1 week → 2 years) from six tables: Actions, Sessions, OEManager Probes, Profiled Requests, Health Scanner Details, and Alerts. Each row shows a live approximate record count, a period selector, and a Truncate button protected by a confirmation modal.

– **Service Actions tab** — audit log of all cleanup operations showing who ran them, which table, what period, how many records were deleted, and a result badge.

– **Scheduler Actions tab** — history of all scheduler operations (start, stop, remove, stop all, restart) with date, user, job name, parameters, and colour-coded result badges.

– **Scheduler Jobs grid enriched** — each job now shows type (interval vs cron), interval in seconds, and cron expression alongside the existing ID and status.

– **Scheduler restart modal** — after a successful restart an info modal confirms completion. The Jobs grid shows a loading spinner during the restart.

– **What’s New modal** — a new menu item in the user dropdown opens this release notes panel, always accessible from any page.

– **Settings — Intelligence tab** — new dedicated tab in Settings for independently enabling or disabling Anomaly Detection, Predictive Recycle, and PASOE Ranking; each toggle includes contextual help explaining the algorithm and requirements.

– **Setup — Session memory threshold** — new per-PASOE integer field on the ABL Sessions tab of the PAS instance edit form; sets the memory threshold (MB) used for predictive recycle analysis. Set to 0 to exclude a PASOE from recycle recommendations entirely.

– **Intelligence page** — new dedicated navigation section for AI-driven analysis with six tabs, a shared filter bar (time range, customer, PASOE), auto-refresh, and cross-navigation to related health pages.

– **Intelligence — Anomaly Detection tab** — list of anomaly alerts with timestamps, probe names, and deviation metrics; click any row to drill down to the affected PASOE, host, or environment.

– **Intelligence — Anomaly Detection Timeline tab** — same anomaly alerts displayed in a vertical timeline with colour-coded severity dots.

– **Intelligence — Anomaly Detection Graph tab** — line chart tracking probe deviation in σ units over time with dashed reference lines at 2σ (marginal) and 3σ (critical).

– **Intelligence — PASOE Ranking tab** — horizontal stacked bar chart ranking all PASoEs by total anomaly count split into critical (red) and marginal (orange) bars; hover for average and peak σ breakdown.

– **Intelligence — Predictive Recycle tab** — list of recycle recommendation alerts showing application memory usage, projected memory, and estimated time to threshold.

– **Intelligence — Predictive Recycle Timeline tab** — same recycle recommendations displayed in a vertical timeline with colour-coded urgency dots.

– **Scheduler — Agent Maintenance Insights tab** — analysis of agent recycle and ABL session cleanup activity showing frequency by hour, day of week, and per PASOE/app, with suggested optimal maintenance windows.

– **Setup — Agents Recycle tab** — new tab on the PAS instance edit form for configuring per-PASOE agent recycle settings.

– **New Dashboard** — redesigned dashboard providing an improved overview of PASOE instances and their health status.

– **Dashboard — ABL Sessions chart** — click the session pie chart on any ABL app card to open a per-agent active/idle sessions timeline.

– **Dashboard — Active Sessions % chart** — click the sparkline trend on an ABL app card to open a dual chart showing active session percentage and session totals over time.

– **Dashboard — Agents Memory chart** — click the memory badge on an ABL app card to open a per-agent memory consumption (MB) timeline.

– **Dashboard — Health Probe chart** — click any health bar label or worst-probe badge to open a single probe score trend (0–1 scale) over time.

– **Dashboard — All Probes chart** — click the chart icon next to the Health Scanner header to compare all health probes for a PASOE side by side.

– **Dashboard — Memory vs HTTP Requests chart** — click the total PASOE memory badge to open a dual-axis chart correlating agents memory with HTTP request volume.

– **Dashboard — ABL Sessions vs HTTP Requests chart** — click the bar-chart icon next to a PASOE name to correlate active ABL sessions with HTTP request volume.

– **Dashboard — Memory Trend & Predictive Recycle chart** — click the memory sparkline or a row in the Recycle Recommendations panel to see a projected memory trend line with the recycle threshold.

– **Dashboard — Anomaly Detection chart** — click the 📊 button next to a PASOE Health Scanner header to open a per-probe σ deviation chart with 2σ and 3σ reference lines.

– **Dashboard — Memory Recycle Recommendations panel** — summary panel listing ABL apps with rising memory trends projected to exceed the recycle threshold, sorted by urgency.

– **Dashboard — Anomaly Summary panel** — summary panel listing all probes currently flagged above 2σ; click any row to open that probe’s chart.

## Backend

– **ABL Sessions Cleanup Analysis** — new endpoint aggregating ABL session cleanup activity by hour of day, day of week, and per PASOE/application; provides the data for the cleanup section of the Agent Maintenance Insights tab.

## Docker Compose Stack

– **Docker log rotation** — all Docker stack services now use Docker JSON log rotation (10 MB / 5 files per container), preventing disk exhaustion in long-running deployments.

– **Configurable log level** — backend log verbosity is now controllable via `logLevel` in the app server config or the `LOG_LEVEL` environment variable, changeable without a redeploy. Production default is `error` for minimal output.