Roles & access control¶
Start with the identity doing the work. The person using TrackMe, the owner of a scheduled report, and the account reading a remote system can be three different users. Giving one of them access does not grant it to the others.
This guide explains those boundaries for all tenants, with a setup map for the next-generation components: UAM, Splunk HEC, Splunk S2S, Splunk UF Fleet, Splunk Forwarder Collectors, Cribl Stream and Cribl Edge. Use Roles & access control — in depth for the capability lists, index permissions, examples and checks.
Four questions before granting access¶
Question |
Identity |
Where to grant access |
|---|---|---|
Who may use this tenant? |
The person signed in to TrackMe |
TrackMe capabilities, the tenant’s role lists and access to its TrackMe indexes |
Who runs its schedules? |
The tenant owner: |
On the search tier hosting TrackMe; see The scheduled owner and service account |
Who searches the source data? |
Locally: the scheduled execution identity. Remotely: the Splunk user behind the selected remote account’s token |
On the search tier where the search runs: search capability, source indexes and any component-specific reads |
Who reads the additional API? |
A separate Splunk remote-account user, or a Cribl connection credential |
On the endpoint being read: HEC endpoint, cluster manager, Deployment Server or Cribl Leader |
A remote account has two access checks. On TrackMe, its role list decides who may use it. On the destination, the token user’s capabilities and index permissions decide what it can read. Both must work for scheduled collection; testing the account as a Splunk administrator does not prove that the tenant owner may use it.
The three TrackMe user levels¶
TrackMe ships three roles. You can assign them directly or inherit them into your own
roles; trackme_admin inherits trackme_power, which inherits trackme_user.
Built-in role |
Capability |
What it is for |
|---|---|---|
|
|
Read the tenant and its monitoring results. |
|
|
Operate existing monitoring. In findings workspaces: triage findings, assign them, add notes and perform the actions the component exposes. |
|
|
Create and administer TrackMe content. For a tenant the user administers: source settings, policies, collection and automation settings. |
The capability grants an operation level; the tenant’s user, power and admin role lists grant the scope. A TrackMe administrator of one tenant does not thereby administer every tenant. Splunk administrators retain administrative access.
These are the permissions of TrackMe users. They do not automatically provide a remote Splunk account with cluster-read capabilities, allow a search of internal indexes, or satisfy the requirements of a named scheduled owner.
Important
UAM has a narrower workspace
UAM contains audit and account activity. Its workspace is for tenant administrators
(trackmeadminoperations plus a tenant admin role), or a Splunk administrator with
admin_all_objects. Tenant power users have selected REST read and action access;
read-only tenant users have no UAM access. See UAM: audit access and per-tier inventories.
Choosing the scheduled owner¶
A dedicated service account makes scheduled activity attributable to a named identity.
It is optional for ordinary collection: nobody is also a supported knowledge-object
owner, with execution supplied by Splunk. Scheduled AI Agents require a named account;
nobody and splunk-system-user cannot run them.
For HEC, S2S, UF Fleet, Forwarder Collectors, Cribl Stream and Cribl Edge, the current creation and collection
checks require a named owner to hold the three TrackMe capabilities, schedule_search,
list_settings, list_storage_passwords and admin_all_objects (or be a member of
admin / sc_admin). Search and index access must also be configured. These broader
permissions belong to the scheduled owner, not to every person using its workspace.
UAM’s collection requirements are checked separately. Its identity and audit reads have different local and remote requirements. The general Configuration Guardian owner check is broader than the UAM collection check; see The scheduled owner and service account and UAM: audit access and per-tier inventories before treating them as the same requirement.
What each next-generation component reads¶
Component |
Data searched in Splunk |
Additional API access |
|---|---|---|
|
User, role, knowledge-object and, for a cluster, SHC member reads on each tier |
|
|
Optional accounts on the HEC endpoints’ splunkd management API for configuration comparisons |
|
|
|
|
Forwarding telemetry on its source tier; separate Deployment Server source for
|
Optional Deployment Server API accounts for current server-class membership;
these accounts also need the three |
|
|
Optional collector API accounts, one per collector, for the current list of its
inputs: |
|
Cribl internal metrics in a metric index searchable from the local TrackMe tier; a remote metric-source account is not supported |
Required Cribl connection with read access to Stream inventory, configuration and Leader-proxied Worker status |
|
No Splunk source index required for core API monitoring; optional local System Metrics index for node CPU and memory |
Required Cribl connection with read access to Edge Fleets, Nodes, configuration and Leader-proxied status |
local describes where a search is dispatched, not where events are stored.
A local TrackMe search head can search distributed indexers. Conversely, a remote
Deployment Server account can search indexes held on that server itself.
Tip
Internal indexes need an explicit grant
Access to all ordinary indexes (*) does not include internal indexes. Grant the
source reader the actual indexes it needs, especially UF Fleet’s three _ds*
indexes. The detailed reference includes index-access examples.
Grant tenant access without exposing every tenant¶
Define your user, power and admin roles in Splunk, with the appropriate TrackMe capabilities and application access. SSO group mappings can assign these roles.
Put those roles in the matching RBAC lists of each tenant.
Give users access to that tenant’s TrackMe result indexes. Grant source indexes and remote-account use separately when their investigations need them.
Configure the owner and each source/API account using the component requirements above.
Check a scheduled collection, not only a manual run. The two may execute under different identities. Follow Validate each identity and diagnose the missing layer for a repeatable check.
Warning
Role inheritance can widen tenant scope
A new tenant’s default lists include the built-in TrackMe roles. If your custom role inherits one of them and it remains listed on other tenants, the inherited role can grant access there too. Replace the default tenant lists with your scoped roles, or grant capabilities directly to your custom roles. Include the lower-level capabilities needed by each higher-level role when using direct grants.
System configuration and tenant administration¶
Configuration → System settings and the dedicated account-management pages are Splunk administrator tasks. They are separate from managing a tenant you administer. Cribl credentials are managed under Configuration → Manage Cribl connections; Splunk remote-account credentials under Configuration → Manage Remote deployment accounts.
Set defaults for new tenants under System settings → General → Knowledge objects & RBAC defaults. To change an existing tenant’s roles or owner, use its Virtual Tenant UI; TrackMe updates the tenant and its knowledge-object permissions. Changing defaults alone does not re-permission existing tenants.
See also
Roles & access control — in depth — exact requirements, component details, account examples and diagnosis.
Creating a tenant and Managing tenants — tenant lifecycle.
Configuration — native configuration interfaces.
Remote Splunk deployments and Cribl Stream & Cribl Edge — connection setup.