Settings
Settings is where you run your workspace — the OTLP credential, who’s on the team, what fires and where it goes, appearance, your usage, and your account. Most of it is self-explanatory; this page is the map of what’s where.
It lives under Admin & Governance → Settings, grouped into Project, Integrations, Preferences and Account. Sections that depend on a workspace feature — Alerting, Notifications and Billing — appear only when that feature is on.
Project
Section titled “Project”App Onboarding
Section titled “App Onboarding”Settings → App Onboarding holds your workspace’s Ingest Tokens and the ready-to-paste
exporter config, the same screen the getting-started flow takes you to.
A new workspace that has never sent anything opens on a welcome offering Connect your app (this
screen) or Explore the demo (the OTel Demo workspace); Skip for now goes straight to the
app.
Generate token opens a dialog. Name the token (it starts with a default such as my-app-token,
or deploy-token for a deploy token, numbered if that name is taken) and choose what it’s for: Send telemetry (the default) or Record
deploys, a token your CI uses to record your deploys. The dialog then
shows the token once for you to copy. A telemetry token is also filled into the snippets; a deploy
token can’t send telemetry, so it isn’t.
Tokens belong to the workspace, not to the person who created them, and they don’t expire: a token
works until you revoke it. The token list shows each token’s name, prefix, purpose, status and when it
was last used. Click a name to rename it. Only a workspace owner or admin creates and manages
tokens; a member is told that an owner or admin creates them. While you’re viewing another
workspace, such as the OTel Demo, the card offers Switch to My workspace: one click back to your
own.
The screen runs in three steps. 1. Create an ingest token explains the token and holds the controls above. 2. Send telemetry from your app offers two routes: My app already uses OpenTelemetry (the default) gives Environment Variables, OTel Collector, Node.js and Python snippets, each built with your real endpoint and token; Have my coding agent do it gives an editable prompt, for any language, that your coding agent follows to wire your app.
3. Check that your telemetry arrived checks nothing until you press I’ve started my app — check now (it starts on its own only when your telemetry is already being accepted). It then shows three stages: Accepted (a request carrying your token was accepted), Searchable (your telemetry is stored) and Analysis ready (calls between your services are mapped into critical paths). Once your telemetry is searchable, Open your trace opens the newest trace it found, beside a next step. When nothing arrives it links straight to the connection diagnostics guide.
Members — your team
Section titled “Members — your team”Settings → Members is where you invite teammates, set their roles, and see who’s on the team. Inviting is one line — type an email, pick a role, press Enter — and new people are set up on the spot while existing atrim.ai users join with one click. The full walkthrough, the three roles and what each can do, pending invitations, and switching between workspaces are covered in their own guide.
General
Section titled “General”Settings → General shows a workspace identifier you can copy and a live Platform Status readout of atrim.ai’s own services. Its Danger Zone holds Reset Tenant Data — a one-action wipe of all telemetry and analysis (traces, logs, metrics, topology, baselines, incidents) that keeps your settings, ingest tokens, webhooks and plan. It’s handy right after onboarding with a throwaway app. Only a workspace owner or admin can run it; a member sees why it’s unavailable rather than a dead button.
Integrations
Section titled “Integrations”Both Integrations sections appear once alerting is enabled for your workspace. Together they split one question in two: Alerting decides what fires and how loud; Notifications decides where it goes.
Alerting — what fires and how loud
Section titled “Alerting — what fires and how loud”Settings → Alerting tunes detection: which detectors are on, how loud each one is (its severity), and the thresholds behind it. Detectors are grouped by alarm class — Baseline alarms fire when a signal drifts from its own trailing normal, and Condition alarms fire when a stated condition holds (a traffic onset or drop, a service going absent). Each row shows whether it’s running on the Default or a Custom setting, and you can reset a custom one back to the default in place.
Notifications — where alerts go
Section titled “Notifications — where alerts go”Settings → Notifications owns delivery: webhooks and email. Add Webhook takes a Name, a
URL, an optional signing secret (sent as an X-Atrim-Signature HMAC-SHA256 header so you can
verify deliveries), and a severity filter — Critical / Warning / Info — so a webhook only receives
what you choose. Each webhook can be paused, edited, deleted, and test-fired in place, so you can
confirm it works before an incident depends on it.
Email needs no setup: incident notifications also go to your sign-up address by default.
Preferences
Section titled “Preferences”Appearance
Section titled “Appearance”Settings → Appearance sets the colour scheme — System, Light, Dark, or the branded atrim.ai mode — saved to your account. A Service Topology control here, Show external dependencies, toggles whether uninstrumented peers (databases, caches, queues, third-party APIs) inferred from outbound calls are drawn on the map. Turn it off to see only services that report telemetry of their own.
Training Data
Section titled “Training Data”Settings → Training Data is your workspace’s choice about contributing telemetry examples that atrim.ai uses to develop, test and improve its static detection rules. They are never used to train or fine-tune models. Selected examples are included by default, and a workspace owner or admin can opt out at any time. atrim.ai selects fields and masks identifiers first, but masking doesn’t guarantee anonymity: service names and other content may identify your organization, and personal data may remain. The card records the current choice and when it last changed; members see it read-only. Opting out excludes your workspace from later exports. Examples already in the dataset stay until they’re deleted: email privacy@atrim.ai and they’re deleted within 30 days. Deleting the workspace also deletes them within 30 days.
Account
Section titled “Account”Profile
Section titled “Profile”Settings → Profile shows your email and connected sign-in providers, and can delete your account. This is about you, not the workspace — it travels with you across every workspace you belong to. Deleting your account deletes your workspace and its data only if you’re its only member. If other members remain, your access is removed and the workspace stays.
Settings → Usage is your own metered-usage view, and it’s always available. It shows, per metered action — AI investigations and QuerySpec / analytics queries — what you’ve used this hour against its cap, plus your month-to-date total. Hourly caps protect shared capacity; your usage is recorded whether or not your plan is capped. A separate Record volume card tracks ingested records for the month and only calls out a number once you’re over your plan’s included volume. On Starter and Pro, past your plan's included volume, ingest keeps flowing at full rate — you're prompted to upgrade, your bill doesn't move, and there is no overage billing. Scale's contracted volume carries a per-record rate beyond it. It’s read-only. Prices and included volumes are on Plans and trial.
Billing
Section titled “Billing”Settings → Billing (when enabled) shows your plan or trial and opens the Stripe portal to manage your card, invoices or cancellation — those details deliberately live in Stripe, not here.