> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://developers.alephant.io/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://developers.alephant.io/_mcp/server.

# Alerts

> Monitor budget, quota, usage anomaly, and operational alerts across agents and workflows

Alerts help teams respond before AI usage becomes expensive, unreliable, or hard to explain.

Alephant alerts should point back to the relevant workspace, department, member, agent, Virtual Key, policy, request, or run trace so operators can investigate quickly.

## Alert Types

| Alert type        | Typical trigger                                                                | Common action                                   |
| ----------------- | ------------------------------------------------------------------------------ | ----------------------------------------------- |
| Budget alert      | Workspace, department, agent, member, or Virtual Key budget passes a threshold | Review spend, throttle traffic, or raise budget |
| Usage anomaly     | Request count, token usage, or cost spikes compared with recent history        | Inspect run traces and recent workflow changes  |
| Quota alert       | Monthly recorded-log or usage quota approaches plan limits                     | Upgrade, export, or reduce noisy traffic        |
| Operational alert | Error rate, latency, fallback rate, or cache behavior changes                  | Investigate provider, routing, or policy state  |
| Positive alert    | Cache savings or routing savings reaches a notable value                       | Review optimization impact                      |

## Severity Levels

| Severity | Meaning                   | Example                                                          |
| -------- | ------------------------- | ---------------------------------------------------------------- |
| High     | Needs immediate attention | Budget usage above hard-stop threshold or request failure spike  |
| Medium   | Needs planned review      | Budget usage above warning threshold or unusual request growth   |
| Low      | Informational             | Cache savings, policy change, or successful routing optimization |

## Budget Alerts

Budget alerts can apply to:

* Workspace budgets
* Department budgets
* Agent budgets
* Member budgets
* Virtual Key budgets
* Paid endpoint or workflow spend limits

Common thresholds are warning, critical, and hard stop. A warning should notify operators; a hard stop should block or pause the relevant traffic according to policy.

## Anomaly Alerts

Usage anomaly alerts are useful when an agent or workflow changes behavior unexpectedly:

* Request volume increases sharply.
* Token usage per request grows.
* A workflow starts retrying repeatedly.
* A fallback route becomes the primary route.
* A paid endpoint begins losing margin per call.

The investigation path should start from the alert and move into logs, run traces, analytics, and policy history.

## Related Pages

* [Cost Analytics](/docs/overview/fin-ops-budget/cost-analytics)
* [Budget Management](/docs/overview/fin-ops-budget/budget-management)
* [Policies & Rules](/docs/overview/security-compliance/policies-rules)
* [Agent Run Tracing](/docs/overview/core-concepts/agent-run-tracing)