TrackMe Capability × Component Matrix¶
What this page is
A flat, exhaustive map of every cross-cutting TrackMe capability
against the seven monitoring components (DSM, DHM, MHM,
VOL, FLX, FQM, WLK). For each (capability × component)
intersection it records:
whether the capability applies to that component,
the use case unlocked by turning it on (one line each),
the configuration touchpoint (UI page, UCC tab, KV collection,
vtenant_accountfield).
Use this page when you know which components a tenant has enabled and want to audit which cross-cutting features are wired in for those components. It pairs naturally with TrackMe Use Cases — Organised by User Intent (which asks the inverse question: given a goal, which component delivers?).
Legend¶
In every applicability table:
Y— fully supported on that component.P— partial support; the cell footnote explains the caveat.—— not applicable on that component (see What does not apply, and why at the bottom of this page).
The seven components — recap¶
For the full description see TrackMe Use Cases Referential — Overview. Short form:
Code |
Full name |
Tracks |
|---|---|---|
|
Data Source Monitoring |
|
|
Data Host Monitoring |
Host × sourcetype activity. 2.3.23+ supports |
|
Metric Host Monitoring |
Hosts × metric category, |
|
Volume Outliers |
Licensed volume per index from the license usage logs (5-minute buckets, incremental). ML outliers (drops / spikes, direction per index), inactivity, daily volume, trend, month projection. 2.4.18+; see VOL — Volume Outliers. |
|
Flex Objects |
Custom SPL turned into tracked entities. 65 pre-built templates. |
|
Field Quality Monitoring |
Field parsing quality, CIM compliance. |
|
WorkLoad Knowledge |
Scheduled-search execution health, versioning, orphan, app enablement. |
User Activity Monitoring (UAM, Beta, 2.4.18+) sits outside this
matrix on purpose: it is a dedicated tenant type with no entities, so the
state, scoring and alerting capabilities below do not apply to it. It
raises findings from a policy catalogue instead, with its own
investigation workflow and notifications. See UAM — User Activity Monitoring (Beta).
State, scoring, alerting¶
These capabilities are largely uniform across components — TrackMe was designed so the alerting and scoring experience is the same regardless of what is being monitored.
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Hybrid impact scoring ( |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Same additive formula across all seven; per-component weight names. |
4-state model (green / orange / red / blue) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Same machine across all seven components. |
Stateful alerting (open / update / close) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Common |
Acknowledgments (sticky / unsticky) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Common ACK collection. |
Auto-ACK alert action |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Triggered on any saved-search firing. |
Notable events alert action |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
One event per fire; not de-duplicated by TrackMe. |
Smart Status investigations |
Y |
Y |
P |
P |
Y |
P |
Y |
DSM / DHM / FLX / WLK have explicit per-component investigation templates; MHM, VOL and FQM inherit the generic outliers template. |
Disruption queue (grace period) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Identical code path across components. |
Email delivery (SMTP) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Same SMTP account; same stateful helper. |
Embedded 24h charts in alert emails |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
|
AI status reports in alert emails |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
|
Free-style REST call alert action |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Targets any |
Use case unlocked (one line per capability):
Hybrid impact scoring — Each anomaly type contributes additively, so one critical breach trips red immediately while several minor ones accumulate into orange before tripping critical.
4-state model — One uniform “what colour is my donut” semantic across every component, every UI surface, every downstream consumer.
Stateful alerting — One email per real state transition, not one email per evaluation cycle; opens / updates / closes thread together.
Acknowledgments — Silence known issues without losing the underlying state signal. Sticky vs unsticky lets you pick auto-expiry.
Auto-ACK — Pre-emptively suppress notifications for entities matching a pattern (“ACK for 2 h on any red in this group”).
Notable events — Feed TrackMe state into Splunk ES / correlation searches / incident platforms via a notable-index event.
Smart Status — When an entity goes red, an investigation search runs once / 24 h and surfaces likely root-cause hypotheses.
Disruption queue — Fewer false-positive pages from 30-second blips; alerts only fire after the configured grace duration of continuous red.
Email delivery — Notifications land in inboxes correctly threaded with embedded charts, no external image fetches.
AI status reports — Recipient sees a 2 – 3 paragraph contextual investigation directly in the email body, not just a link.
Free-style REST call alert action — Wire any TrackMe REST mutation (entity CRUD, mode flip, ACK bulk-op) into a saved-search firing without writing a new alert action.
Classification and protection (policies)¶
This is the largest cluster of cross-cutting features. Several of them are DSM / DHM specific because they relate to delay / latency thresholds.
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Priority policies (regex / lookup / SPL) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Per-component policy collection. |
Tag policies (regex / lookup / SPL) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Multiple matches union. |
SLA policies (regex / lookup / SPL) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
“Last wins” when multiple policies match. |
Labels (catalog-backed coloured chips) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Single catalog per tenant. |
Notes (Markdown annotations) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
One tenant collection. |
Variable delay (time-slot thresholds) |
Y |
Y |
— |
— |
— |
— |
— |
DSM / DHM only. |
Lagging classes (named delay / lag rule sets) |
Y |
Y |
— |
— |
— |
— |
— |
DSM / DHM only. |
Adaptive delay (automated threshold recompute) |
Y |
Y |
— |
— |
— |
— |
— |
DSM / DHM only. 2.3.23+ also recomputes variable-delay slots. |
Variable dynamic thresholds (time-slot KPI thresholds) |
— |
— |
— |
— |
Y |
— |
— |
FLX only. Different mechanism from DSM/DHM variable delay. |
WLK tenant-default threshold row |
— |
— |
— |
— |
— |
— |
Y |
WLK-only fleet-wide fallback ( |
Logical groups (blue protection) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Cross-component membership supported. |
Blocklist / Allowlist (regex) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Per-component allowlist collection holding both rule types. |
Monitoring-time policy |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Tenant default + per-entity override. |
Use case unlocked:
Priority policies — Don’t manually set priority on 5,000 entities; a regex on the index name covers them.
Tag policies — Routing labels (
team:auth,env:prod) attach automatically and survive entity recreation.SLA policies — Cleanly bucket entities into platinum / gold / silver uptime tiers per regex or lookup.
Labels — Visual lifecycle workflow (
under-review→in-progress→resolved) with curated colours.Notes — Persist investigation context for the next operator without polluting structured fields.
Variable delay — Don’t alert on the expected nightly-batch window; business-hours threshold stays tight.
Lagging classes — Manage thresholds once for 40 high-frequency feeds by editing one class instead of 40 entities.
Adaptive delay — Persistently-red
delay_threshold_breachedentities get their thresholds auto-recomputed from history.Variable dynamic thresholds (FLX) — Tighter KPI thresholds during business hours, looser overnight, on the same metric.
WLK tenant-default threshold row — Manage skip% / error count / delay thresholds for the whole tenant from one record; per-entity overrides as exceptions.
Logical groups — Don’t alert when only 1 of 100 hosts is down; the cluster’s green % protects struggling members.
Blocklist / Allowlist — Scope discovery to
index=prod_*only; never tracksourcetype=.*:debug.Monitoring-time policy — Auto-pause monitoring outside business hours; entities go orange (silent) instead of false-positive red.
Discovery, sampling, and quality¶
These are mostly component-specific signals — DSM-only sampling features, FQM-only CIM templates, WLK-only signals, and DHM-only extras.
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Data sampling (regex format models) |
Y |
— |
— |
— |
— |
— |
— |
DSM-only — needs raw events. |
Elastic sources (external collectors) |
Y |
— |
— |
— |
— |
— |
— |
DSM-only. |
Lookups monitoring ( |
Y |
— |
— |
— |
— |
— |
— |
DSM-only; requires |
CIM-aware field-quality templates |
— |
— |
— |
— |
— |
Y |
— |
FQM-only. |
Sourcetype-explosion safeguard |
Y |
— |
— |
— |
— |
— |
— |
DSM-only entity-creation hot path. |
WLK orphan / versioning / app enablement |
— |
— |
— |
— |
— |
— |
Y |
WLK-only signals. |
DHM |
— |
Y |
— |
— |
— |
— |
— |
DHM hybrid trackers, 2.3.23+, Custom break-by mode only. |
History backfill at creation |
— |
— |
— |
Y |
— |
— |
— |
VOL-only: the license-usage history (default 30 days, 0–90) is replayed automatically so outliers models and the cost trend are usable on day one. |
Inactivity policy (static or day-of-week / hour-of-day) |
— |
— |
— |
Y |
— |
— |
— |
VOL-only: tenant-level policy, overridable per index; an index that stops consuming license turns red. |
Outlier detection direction per entity (drops / spikes / both) |
— |
— |
— |
Y |
— |
— |
— |
VOL-only: chosen per index. |
30-day trend, month projection, Global license usage analytics |
— |
— |
— |
Y |
— |
— |
— |
VOL-only: per-index trend and projection; the Global license usage (VOL) Tenant Home tab covers key figures, insights, top consumers, movers, pools and quotas. |
ML and advanced trackers¶
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Native ML outlier detection |
Y |
P |
P |
Y |
Y |
P |
P |
Recommended for DSM and FLX (natural fit); the primary state
signal for VOL (enabled by default, default model on the rolling
24 h volume, own |
Hybrid trackers (custom per-entity SPL check) |
Y |
Y |
Y |
— |
Y |
Y |
Y |
One registry per component. Enterprise caps at 8 per tenant; Foundation rejects; Unlimited uncapped. Not applicable to VOL (its single collector is the tracker). |
Replica trackers (one-way entity replication) |
Y |
Y |
Y |
— |
Y |
Y |
Y |
Source and target component must match. Local decision maker does not re-evaluate replicated entities. |
Use case unlocked:
Native ML outlier detection — Get alerted on volume / KPI anomalies without writing thresholds; the model learns the normal pattern with optional weekly seasonality.
Hybrid trackers — Add a custom SPL check that contributes to an entity’s score; useful when the default tracker cannot reach the data or you need a domain-specific check on top of the built-in tracker.
Replica trackers — Single pane of glass in a central TrackMe across regional TrackMes without running searches on each for every UI refresh.
Integrations¶
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Remote Splunk accounts ( |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Per-tenant |
AI Assistant chat (per-entity context) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Entity-level context types exist for all seven components. |
AI ML Advisor (interactive + automated batch) |
Y |
Y |
— |
Y |
Y |
Y |
Y |
Inspects / remediates native ML outlier models; the automated batch covers DSM, DHM, VOL, FLX, FQM, WLK. |
AI Components Advisor (Feed Lifecycle / FLX Threshold / FQM / Component Health) |
Y |
Y |
Y |
— |
Y |
Y |
Y |
No Components Advisor covers VOL; the ML Advisor and the Concierge do. |
Topology Studio views (entity nodes) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Any entity from any tenant and component can be placed on a view. |
Tenant Home topology graph (link / network) |
Y |
Y |
Y |
— |
Y |
Y |
Y |
VOL has one entity per index — no hierarchy to draw. |
Splunk SOAR push (containers + enrichment) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Component-agnostic; works on any rows producing entities. |
Use case unlocked:
Remote Splunk accounts — Monitor any number of remote Splunk deployments from a central TrackMe with HA URLs and automatic token rotation.
AI Assistant chat — Ask “why is this entity red?” on the Tenant Home page and get a context-aware LLM answer that knows the entity’s state, policies, recent ACK history, and active Guardian alerts.
Splunk SOAR push — Auto-create SOAR containers from TrackMe alerts for runbook-driven remediation; broker-failover safe.
Multi-tenancy and operations¶
Cross-cutting capabilities that sit above components. They apply uniformly to every component a tenant has enabled.
Capability |
DSM |
DHM |
MHM |
VOL |
FLX |
FQM |
WLK |
Notes |
|---|---|---|---|---|---|---|---|---|
Virtual Tenants (isolation, RBAC, indexes) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Components are enabled per tenant via |
Virtual Groups (read-only cross-tenant cards) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
|
RBAC 3-tier (user / power / admin) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Same model regardless of component. |
Maintenance mode (runtime alert silencing) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Global or per-tenant; suppresses alerting-path actions. |
Maintenance KDB (windows, SLA exclusion) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Scope cascade global / tenant / component / entity. |
Bank holidays (SLA exclusion) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Tenant opt-in. |
Backup & restore (multi-archive 3.0.0) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Tenant archive carries every component’s tenant collections. |
Configuration Guardian |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
All registered checks are system- or tenant-scoped, not component-specific. |
Schema migrations (per-tenant version) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Each migration may touch any component’s collections. |
License enforcement (read-only, edition gates) |
Y |
Y |
Y |
Y |
Y |
Y |
Y |
Foundation gates FLX, FQM and WLK creation (VOL is included in Foundation); Enterprise caps hybrid trackers per tenant. |
Configuration touchpoints¶
Where each capability is configured. UCC refers to the
Configuration page tabs driven by globalConfig.json;
vtenant_account is the JSON blob inside the tenant record.
Capability |
Touchpoint |
|---|---|
Hybrid impact scoring weights |
Tenant-level: |
Stateful alerting |
Alert-action parameters on the saved search (Alert Editor); state
in |
Acknowledgments |
Tenant Home entity → Ack modal; REST |
Auto-ACK / Notable / Free-style REST call / Smart Status |
Alert-action parameters on the firing saved search. |
Email delivery |
UCC |
AI status reports in alert emails |
Alert-action |
Disruption queue |
System default in |
Priority / Tag / SLA policies |
Tenant Home per-component “Policies” modal; REST
|
SLA classes |
System-level |
Labels (catalog + assignments) |
Tenant Home “Manage Labels” modal; REST |
Notes |
Entity overview “Notes” panel; REST |
Variable delay (DSM / DHM) |
DSM / DHM entity → “Variable delay threshold” modal; REST
|
Lagging classes (DSM / DHM) |
Tenant Home DSM / DHM “Lagging classes” modal; REST
|
Adaptive delay |
Per-entity |
FLX / FQM / WLK dynamic thresholds |
Tenant Home per-component “Thresholds” modal; REST
|
Logical groups |
Tenant Home per-component “Logical groups” modal; REST
|
Blocklist / Allowlist |
Per-component wizard / feed configuration; REST
|
Monitoring-time policy |
|
Data sampling (DSM) |
DSM tab → “Data sampling” modal; REST
|
Elastic sources (DSM) |
DSM “Elastic sources” wizard; REST
|
Lookups monitoring (DSM) |
Hybrid Trackers create wizard, Step 1 |
WLK orphan / versioning / app enablement |
WLK wizard + per-app
|
Sourcetype-explosion safeguard |
|
Native ML outlier detection |
Tenant Home per-component “Outliers rules” modal; tenant-level
|
VOL detection direction, inactivity policy, backfill |
Tenant Home VOL tab (per-index direction and inactivity override,
backfill control) and the Global license usage (VOL) tab; tenant
inactivity policy and |
Hybrid trackers |
Tenant Home per-component “Hybrid Trackers” modal / wizard; REST
|
Replica trackers |
Virtual Tenants “Replica Wizard”; REST
|
Remote Splunk accounts |
UCC |
AI Assistant (provider config) |
UCC |
AI Assistant (per-tenant) |
|
Splunk SOAR |
|
Virtual Tenants |
UCC |
Virtual Groups |
Virtual Tenants page Create Virtual Group wizard; REST
|
RBAC (per-tenant roles) |
Splunk roles
( |
Maintenance mode |
Maintenance Mode page; REST |
Maintenance KDB |
Maintenance KDB page; REST |
Bank holidays |
Manage Bank Holidays page; REST |
Backup & restore |
Backup & Restore page; REST
|
Configuration Guardian |
Surfaced as toasts on Virtual Tenants page; REST
|
Schema migrations |
Per-tenant |
License |
License page; REST |
What does not apply, and why¶
DSM-only capabilities¶
Data sampling — Inspects raw events; needs
index + sourcetype + _raw. DHM works off ingest metadata (no raw text), MHM is metric data, FLX / FQM / WLK build their own metric shape.Elastic sources — A DSM entity is the right granularity for “is this external collector producing data?”.
Lookups monitoring — Lookup files / KVstore collections are configuration artefacts, not data feeds; DSM is the closest model. Operates with neutralised latency (no
_indextimeexists for lookups) and a single delay-threshold alerting hook.Sourcetype-explosion safeguard — Bulk DSM creation discovers an
(index, sourcetype)matrix that can blow up if a pipeline mis-routes events. Other components do not pivot on this discovery axis.
DSM + DHM only (not MHM / VOL / FLX / FQM / WLK)¶
Variable delay — Day-of-week × hour-of-day thresholds make sense for data feeds and per-host streams that have legitimate cadence variation. MHM uses static thresholds.
Lagging classes — Same rationale; only meaningful for delay / latency-based components.
Adaptive delay — Recomputes
data_max_delay_allowed(and variable-delay slot thresholds since 2.3.23) from historical metrics. Only meaningful for components with a singledelayaxis to recompute.
WLK-only signals¶
Orphan, versioning, app enablement — They describe properties of a Splunk scheduled search and its hosting app. Not applicable to data feeds, hosts, custom SPL trackers, or field-quality entities.
Tenant-default threshold row (``object_id == “default”``) — WLK-specific fallback for “manage thresholds once for the whole tenant, override per entity as exceptions”.
FQM-only¶
CIM-aware templates — FQM is purpose-built for “is this CIM data-model field being parsed correctly?”. The other components do not evaluate per-field coverage / success metrics.
DHM-only (2.3.23+)¶
``breakby_extra_fields`` — Extra metadata dimensions on top of
(index, sourcetype)for DHM hybrid trackers in Custom break-by mode. Specific to how DHM groups per-host activity.
FLX-only¶
Variable dynamic thresholds (time-slot KPI thresholds) — FLX supports day-of-week × hour-of-day slots on its dynamic threshold evaluator. This is a different mechanism from DSM / DHM variable delay — variable delay applies to feed delay thresholds, variable dynamic thresholds apply to any FLX KPI.
VOL-only¶
History backfill, inactivity policy, detection direction, month projection, Global license usage analytics — VOL tracks one signal (licensed volume per index, from the license usage logs) whose state is the ML outlier verdict plus inactivity. There are no delay / latency thresholds, no dynamic thresholds and no raw events to sample, so none of the feed- or threshold-specific capabilities apply; the VOL-specific rows above exist because that signal needs a history to learn from (backfill), a silence rule (inactivity policy) and a cost reading (trend, projection, global analytics).
No hybrid or replica trackers — the VOL tracker is a single incremental collector per tenant; there is no per-entity SPL to customise and no replica path.
No Tenant Home topology graph — one entity per index leaves no hierarchy to draw; VOL entities are placed in Topology Studio views like any other component.
Outlier-detection recommendation¶
DSM and FLX are recommended (event volume / numeric KPIs are natural density-function candidates); on VOL outliers are the point of the component and are on by default. DHM is selective (“per-host volume can be noisy”). MHM, FQM, WLK are not recommended:
MHM — Metric streams are too sparse or too categorical for density modelling.
FQM — Field-quality metrics are slow-moving and regime-dependent; thresholds are the right tool.
WLK — Scheduler metrics are discrete (skip count, error count); again, thresholds are better.
Smart Status template coverage¶
Smart Status investigation templates are documented for DSM, DHM, FLX,
WLK anomalies (delay, future data, latency, hosts dcount, inactivity,
threshold, skip, error, orphan). MHM and FQM fall through to a generic
outliers template — marked P in the matrix.
Replica trackers and the decision maker¶
Replica trackers replicate an entity set but the local decision maker
does not re-evaluate them — it trusts the source’s state. So while the
matrix shows Y across the six components that support them (replica
trackers work on any component except VOL), the surrounding capabilities (logical groups,
disruption queue, policies) only apply to non-replicated entities
in that component; on replicated entities they are inert by design.
License-edition gating¶
Foundation edition supports DSM / DHM / MHM / VOL only; FLX, FQM and WLK rows in the matrix are still
Y(the capability exists), but Foundation tenants cannot create FLX / FQM / WLK entities.Enterprise caps hybrid trackers at 8 per tenant; Foundation rejects hybrid-tracker creation; Unlimited is uncapped.
Read-only mode (license expired) disables creation / editing across all seven components without removing the capabilities themselves.