> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sureops.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Budgets & cost
> Track LLM spend per environment, set budget limits, and receive cost alerts before you overspend
Every incident sureops resolves consumes LLM tokens — for diagnosis, root cause analysis, recommendations, and post-closure summarization. The Budgets & Cost dashboard gives you visibility into where that spend is going and lets you set limits so you are never surprised at the end of a billing period.
***
## What sureops tracks
sureops meters token usage at the agent level and aggregates it up to environment and organization:
| Metric | Where it appears |
| --------------------------------------------- | ------------------------------------------ |
| Tokens used per incident | Incident detail → AI Analysis tab |
| Tokens used per agent node | Incident detail → Pipeline Analytics |
| Daily / weekly / monthly spend by environment | Settings → Budgets |
| Cumulative spend against budget limit | Settings → Budgets → gauge per environment |
| Cost trend over time | Settings → Budgets → chart view |
Spend is calculated using the list price for the configured model per environment. If you supply your own API key (BYOK), the cost shown is based on the same list price — sureops does not have visibility into any negotiated rates you may have with your model provider.
***
## Setting a budget limit
Budget limits are per-environment and are configured by Admins.
Navigate to **Settings → Budgets**. You will see a card for each environment in your organization with current spend and limit (if set).
Choose the environment you want to limit. Click **Edit Budget** on its card.
Enter the monthly spend limit in USD. You can also set a **warning threshold** — a percentage of the limit at which sureops sends an alert before the limit is reached (default: 80%).
Select what happens when the limit is reached:
* **Alert only** — sureops sends a notification but continues operating. Useful for getting baselining data before enforcing hard limits.
* **Pause new incidents** — new incidents can still be opened and tracked, but AI agent pipelines will not run until the budget is reset or increased.
* **Pause all AI** — AI pipelines stop for both new and in-progress incidents. Existing incidents remain open for manual response.
**Pause all AI** stops AI pipelines mid-flight. Any incident currently in an active agent stage will halt at its next safe checkpoint. Choose this option only if you are comfortable with a manual handoff for in-flight incidents.
Click **Save**. The budget limit takes effect immediately. The card updates to show the new limit.
***
## Budget alerts
sureops sends budget alerts via your configured notification channels (email, Slack, in-app). There are two alert types:
| Alert | When it fires |
| ------------------------- | ----------------------------------------------------------------------- |
| Warning threshold reached | When spend reaches the percentage you configured (default 80% of limit) |
| Budget limit reached | When spend hits 100% of the limit |
Alerts go to all **Admins** and **Owners** in the organization. You can add additional notification targets in **Settings → Notifications**.
***
## Resetting or adjusting a budget
Budgets reset automatically on the first of each calendar month. If you need to:
* **Increase the limit mid-month** — edit the budget limit at any time; the new limit takes effect immediately
* **Reset manually** — click **Reset this month** on the environment's budget card; this zeroes the spend counter without affecting historical data
* **Remove a limit** — set the limit to blank or zero to remove all enforcement (alerts still work if a warning threshold is configured)
***
## Cost dashboard
The cost dashboard at **Settings → Budgets → Cost breakdown** shows:
Side-by-side spend comparison across your environments. Identify which environment is driving the most AI usage.
Breakdown by agent type — diagnosis specialist typically accounts for the majority of spend; post-closure summarization is the second-largest contributor.
Top incidents by token spend. Long diagnosis loops and high-severity incidents with many tool calls tend to dominate the list.
30-day rolling spend chart. Useful for spotting anomalies — a sudden spike often correlates with a high-severity incident or a misconfigured agent skill that triggers excessive tool calls.
***
## Per-incident pipeline analytics
For a detailed cost breakdown of a single incident, open the incident detail page and navigate to **Pipeline Analytics**. This view shows:
* Wall-clock time per agent stage
* LLM calls per stage — model used, token count, latency
* Total token spend for the full incident
Pipeline Analytics is useful for tuning: if you see a diagnosis loop that ran 15 LLM calls on a simple incident, that points to a skill or tool configuration worth tightening.
***
## Bring your own key (BYOK)
If your organization supplies its own LLM API keys, spend tracking uses the same token counts but the list-price cost shown in the dashboard may not match your actual invoiced rate. sureops does not receive any billing data from your LLM provider — it only knows what it sent and received.
BYOK keys are configured per-environment in **Settings → AI Agents → LLM Provider**.
***
## Frequently asked questions
**Does the budget limit count toward my sureops subscription cost?**
No. Budget limits track LLM token spend on the model provider's API — not sureops subscription charges. The two are billed independently.
**What model costs are included?**
All models configured for your environments — across all agent types (commander, diagnosis, resolution, verification, ECHO chat). If an agent is configured to use a specific model override, that model's pricing is used for that agent's calls.
**Can I see spend broken down by team member?**
Not at the per-user level. Spend is tracked at the incident and agent level. A per-member view is on the roadmap.
**Is historical cost data retained?**
Yes, indefinitely. Monthly summaries and per-incident detail are stored as long as your organization is active.
***
## Related
* [Notifications](/next/administration/notifications) — configure where budget alerts are sent
* [Roles & Capabilities](/next/administration/roles-and-capabilities) — Admin role required to edit budget limits