Add identity tags

By default, Trove tags every signal with its source harness (harness.id, harness.name) but not with who produced it. Identity tagging adds user.name and user.email as resource attributes on every span, metric, and log Trove forwards, so your backend can group activity by person — which turns “spend by tool” into “spend by person and tool,” and makes dead-seat detection possible.

These are resource attributes: set once per machine, attached to every signal at the source. You don’t tag individual events — Trove stamps the whole stream.

Set your identity

1. Open Trove and go to the Settings tab.

   Note

   Screenshot coming soon.

2. Enter your name in the user.name field and your email in the user.email field. Use the same email your license roster uses, so a person’s activity in Trove lines up with the seat they hold.

3. Save. Trove writes the values to the single versioned state.json, regenerates the collector config so the new resource attributes are applied, and reloads (a brief blip). From the next batch onward, every signal carries your identity.

Tip

These values are plain attribution metadata, not credentials — they live in state.json, not the keychain. (Backend ingest keys are the secrets, and those always go to the OS keychain.) Use the email that matches your SSO / license records so backend joins are clean.

What the tags look like at your backend

Once set, every signal Trove forwards carries the full resource-attribute set:

service.name   = <harness service name>
harness.id     = claude-code
harness.name   = Claude Code
user.name      = Ada Lovelace
user.email     = ada@example.com
trove.source   = trove

Because identity is a resource attribute, it’s available on Tier A metrics, Tier B native metrics, traces, and logs alike — so you can group any of them by user.email in your backend.

What this unlocks

• Per-user dashboards. Group trove.harness.events, trove.harness.tokens, or trove.harness.cost.usd by user.email to see who’s doing what, with which tool.
• Team rollups across mixed harnesses. Since user.email is consistent regardless of which harness emitted, one panel can total a person’s activity across Claude Code, Cursor, Copilot CLI, and the rest.
• Dead-seat detection. Comparing the set of active user.email values against your license roster is exactly how you find paid seats with zero activity. See Normalize costs and find dead seats.

Caution

Identity tags are per-machine: each person sets their own in their own Trove install. For a complete team view, roll this out so everyone has identity configured — anyone who hasn’t set it will still emit telemetry, just without the user.* attributes, and will be invisible to per-user queries. Trove keeps this data local: the attributes are forwarded only to the backend you configured, never to Trove or Intevity.

Normalize costs & find dead seats Put user.email to work spotting unused paid licenses.

Customize mappings Shape the Tier A metrics these identities attach to.

Connect your first backend Have a destination ready to receive the tagged stream.

Settings reference Auto-update and identity options in one place.