Overview

What Topology Studio is

TrackMe knows a great deal about your environment: every monitored entity, its health, its priority, its metrics. Topology Studio lets you arrange that knowledge the way you think about your environment — as a dependency map. A payment flow, a datacenter, an ingestion pipeline, a business service: you decide what belongs on the map and how it connects, and TrackMe keeps the map alive.

A finished topology view — a service map with entity nodes, aggregate groups and edges, colored by live status

A topology view is:

  • Built from live health. Health nodes represent monitored entities, scoped aggregates, auto aggregates of connected nodes, or saved views rolled up as one node. Labels and decorations add context without health calculations. When an entity turns red, so does its node.

  • Authored or generated from Tenant Home. Place nodes, draw edges and choose the icons and layout yourself, or, since TrackMe 2.4.16, generate a live preview from a Virtual Tenant and save it as your starting point. Refine the map to reflect your understanding of dependencies.

  • Shared. Views are stored centrally and shared under role-based access control, with separate read and write permissions — author once, consume everywhere, including full-screen on a wall of screens.

  • Alerting-capable. Any view can carry Topology Alerts — a state-aware engine that correlates the health of the map and notifies on transitions, with the rendered topology image embedded in the notification email.

  • AI-integrated. The AI Assistant understands your views and alerts in context — and can create or modify topologies for you through consented, audited actions.

Core concepts

Views

A view is one topology map: an identifier (view_id, lowercase, up to 40 characters), a display alias, an optional description and category (views sharing a category are grouped together in the list), the canvas content itself, and the role-based access lists. Views are global objects — they are not owned by a tenant, they reference entities across tenants.

Node types

Five kinds of items live on the canvas:

  • Entity nodes — a single monitored entity, referenced by tenant, component and identity. The node carries the entity’s live state color, and up to five KPIs displayed on the canvas (delay, latency, volume, skip rate… — or, for Flex Objects, FQM and WorkLoad, any of the entity’s own metrics).

  • Scoped aggregate nodes — a live roll-up of many entities: one or more tenant/component scopes plus an optional filter (the same filter grammar as Virtual Groups). The node folds to the worst state of everything it matches and shows per-state counts — or, by opt-in, classifies by a healthy-percentage policy with your own thresholds. Aggregates can also auto-populate: their current members appear as linked satellite nodes around them, following membership dynamically.

  • Auto aggregates (TrackMe 2.4.18+) — health parents whose membership comes from incoming connections to entities, aggregates or view nodes. Children keep their scopes and policies; the parent follows their worst effective state and counts unique underlying entities for its KPI. No duplicated filters or independent parent policy are needed. See Auto aggregates for creation, label conversion and nesting.

  • View nodes (TrackMe 2.4.15+) — a reference to another saved topology view, placed on this canvas as a single health node. The referenced view keeps ownership of its graph, its aggregate policies and its own alerts; the parent stores only the identifier, never a copy, so saving the child is reflected at the parent’s next refresh. The node folds to the worst effective state of the child’s health nodes — nested views and impact propagation included — while its KPI counts the unique underlying entities across the child’s aggregates, direct entities and nested views (an orange child with 12 green and 4 orange entities reads 75% healthy, 16 entities and stays orange). Open the referenced view in a new tab from the node’s inspector: the parent canvas, viewport and unsaved edits stay put. Views inside views turn a map into a map of maps — an estate modelled as layered, independently owned views with one roll-up. A view node is a first-class alertable item, and it travels with history, duplication and export like any other node.

  • Label nodes — free-text annotations with color control, for naming zones, tiers or flows on the map.

Five canvas items: entity, scoped aggregate, auto aggregate, view and label A parent view holds a label and three health nodes contributing to an auto aggregate through incoming links. The auto aggregate combines their effective health without a scope or filter. A view node references a separately owned child view by identifier only. Its state follows the child's health nodes, its KPI counts unique underlying entities, and permissions are checked at every hop. Five canvas items — health from data, connections or saved views Auto aggregates combine incoming health nodes; scoped aggregates select entities; view nodes reference maps; labels annotate. Parent view · production-platform authored here — child-to-parent links define auto-aggregate membership Web tier LABEL AUTO AGGREGATE Service — no scope or filter ENTITY docs.trackme.com LATENCY 120 ms SCOPED Web servers HEALTHY · SCOPE + FILTER 100% · 12 entities VIEW Payments HEALTHY · UNDERLYING ENTITIES 75% · 16 entities The auto parent follows its worst child (orange); the label contributes no health. The view node is orange because its child is — it never copies the child's graph. stores the view_id only opens in a new tab Referenced view · payments owns its graph, aggregate policies and alerts Card processing 12 green · 4 orange api.pay 1 entity, green Fraud checks (nested view) 3 entities, all green worst effective state → orange unique entities → 16 STATE The worst effective state of the child's health nodes, after its aggregate policies and impact propagation — nested views included. Labels contribute nothing. Propagation in the parent can then carry that state further along its edges. KPI Unique underlying entities across the child's aggregates, direct entities and nested views, deduplicated by tenant, component and key. Counted separately from the policy-driven colour: 12 green + 4 orange = 75% healthy, 16 entities, while the view stays orange. GUARDS Your read access to the view and to every tenant it reaches is checked at each hop. Missing and denied targets look the same. A cycle, an exhausted nesting or size budget or an incomplete child read never reports healthy — it fails closed.

Live status

Open a view and it polls live status on a fixed cadence (with an honest countdown, a manual refresh, and a pause/resume toggle). Entity nodes colorize by object_state; scoped aggregates fold their scopes server-side; auto aggregates combine their incoming contributors; view nodes resolve the referenced view server-side on every refresh, checking your read access to the view and to every tenant it reaches at each hop. Nodes whose underlying entity was deleted are flagged as missing (with a one-click cleanup); nodes referencing tenants outside your permissions render neutrally as restricted — see tenant isolation. A referenced view that is missing and one you are not allowed to read look the same, so a guessed identifier reveals nothing; and a reference whose health cannot be fully evaluated — a cycle, an incomplete child read, an exhausted nesting or size budget — never reports healthy. Auto aggregates likewise remain unavailable when their contributors are empty or incomplete, rather than showing a partial roll-up as healthy.

How it compares

Feature

Purpose

Topology Studio

Live maps across tenants and components, authored directly or generated from Tenant Home. Saved views are persistent, shared under RBAC and alerting-capable.

Tenant Home Studio preview (2.4.16+)

A temporary live canvas generated from the current selection, with configurable grouping, filtering, inspection and layout. Save explicitly to keep it as a normal Studio view.

Link graph and Network graph

The existing per-component relationship graphs, available as alternate Tenant Home topology tabs with their detailed breakdowns.

Virtual Groups

Read-only aggregation cards in the Virtual Tenants grid — counts across tenants, no canvas, no edges, no alerting.

Hint

Use the Tenant Home preview to discover structure and start a saved Studio view, the Link graph and Network graph for their component breakdowns, and Virtual Groups for quick cross-tenant counters.

Where to find it

Click Topology Studio in the TrackMe navigation bar.

You can also start from the topology icon in any Virtual Tenant’s Tenant Home and save the generated preview into Studio.

The landing page — the views browser — lists every view your roles can see, with a text search and a category filter, in the layout you prefer: Cards (grouped by category), Compact cards (denser tiles) or List (a sortable table). Your choice is remembered per browser; administrators set the deployment-wide default. From there you create, open, duplicate, export or import views. On a fresh install there is nothing to list yet, so the page invites you to create your first view — or to Import one exported from another instance. See Authoring topologies to build the first one.