Skip to main content
GET
Error
Required Permissions How permissions work

Authorizations

Authorization
string
header
required

Management API authentication for /api/* endpoints. Use the Authorization header with Bearer <token>, where <token> is one of:

  • a Bifrost management API key,
  • a dashboard session token issued by POST /api/session/login,
  • base64 of <admin-username>:<admin-password> (legacy equivalent of BasicAuth).

Virtual keys (sk-bf-*) and the x-api-key header are not accepted on management APIs - the sole exception is GET /api/governance/virtual-keys/quota, which is virtual-key-only.

Authentication alone is not sufficient in Bifrost Enterprise: each operation page shows a Required Permissions table (Resource:Operation, for example Dashboard:View) above its Authorizations section, and the caller's RBAC role or management API key scopes must include what it lists, otherwise the request is rejected with 403 Forbidden.

A local admin — authenticated with the admin password, or any caller on a deployment with dashboard auth disabled — bypasses these checks and can call every management endpoint.

OSS setup lock. On Bifrost OSS, while dashboard auth is not active (no admin account, or auth disabled), every management endpoint except the public ones (/health, /api/version, /api/session/is-auth-enabled, /api/session/login, ...) requires the operator's setup token in the X-Bifrost-Setup-Token header, in place of Authorization. The token is set with setup_token in config.json or the BIFROST_SETUP_TOKEN environment variable. A missing header returns 401, a wrong token 403. The header stops working once dashboard auth is enabled. The dashboard instead trades the token once for an HttpOnly bifrost_setup_session cookie via POST /api/session/setup. See Required permissions for how permissions are derived and which endpoints are exempt.

Query Parameters

providers
string

Comma-separated list of providers to filter by

models
string

Comma-separated list of models to filter by

status
string

Comma-separated list of statuses to filter by

objects
string

Comma-separated list of object types to filter by

selected_key_ids
string

Comma-separated list of selected key IDs to filter by

virtual_key_ids
string

Comma-separated list of virtual key IDs to filter by

project_ids
string

Comma-separated list of project IDs to filter by

routing_rule_ids
string

Comma-separated list of routing rule IDs to filter by

routing_engine_used
string

Comma-separated list of routing engines to filter by (routing-rule, governance, or loadbalancing)

tool_call_names
string

Comma-separated list of function names. Matches requests whose response called any of them, using the persisted tool_call_names metadata on the log row. Names are recorded even when content logging is disabled and stay on the row in hybrid storage mode; the call arguments are not.

complexity_tiers
string

Comma-separated list of routing complexity tiers to filter by (SIMPLE, MEDIUM, COMPLEX, or the legacy REASONING tier recorded on historical rows before it was merged into COMPLEX)

complexity_mechanisms
string

Comma-separated list of complexity decision mechanisms to filter by (semantic, decision, llm, session, or skipped)

session_id
string

Exact Bifrost session ID used for key stickiness and request correlation

start_time
string<date-time>

Start time filter (RFC3339 format)

end_time
string<date-time>

End time filter (RFC3339 format)

min_latency
number

Minimum latency filter

max_latency
number

Maximum latency filter

min_tokens
integer

Minimum tokens filter

max_tokens
integer

Maximum tokens filter

min_cost
number

Minimum cost filter

max_cost
number

Maximum cost filter

missing_cost_only
boolean

Only show logs with missing cost

Search in request/response content

request_id
string

Exact lookup on a log ID (which is the request ID). Takes precedence over the time range — start_time/end_time/period are ignored when it is set. roots_only is ignored too, so an ID naming a fallback child returns that child rather than collapsing it into its root.

compare_to_previous
boolean
default:false

When true, also return the same statistics for the immediately preceding window of equal length, so callers can render change-vs-previous-period. Requires a bounded window (start_time and end_time, or period); an unbounded all-time filter has no preceding window, and the response then omits previous with has_previous_period false. A failure to compute the previous period degrades to has_previous_period false rather than failing the request.

Response

Successful response

Log statistics for the requested window. The current period's fields are at the top level, so a request without compare_to_previous returns exactly the LogStats shape.

total_requests
integer<int64>
required
total_tokens
integer<int64>
required
total_cost
number
required
average_latency
number
required
success_rate
number
required

Percentage of completed provider attempts that succeeded

user_facing_success_rate
number
required

Percentage of user requests that ultimately succeeded, counting fallback chains as one request

user_facing_total_requests
integer<int64>
required

Count of root requests used as the denominator for user_facing_success_rate

cache_hit_rate_total_requests
integer<int64>

Completed requests used as the local-cache hit-rate denominator

direct_cache_hits
integer<int64>

Number of direct local-cache hits

semantic_cache_hits
integer<int64>

Number of semantic local-cache hits

has_previous_period
boolean

True when previous is populated. False when the comparison was not requested, the window is unbounded, or the previous-period query failed.

previous
object

The same statistics for the preceding window of equal length, ending immediately before the requested window starts. Present only when compare_to_previous is true and has_previous_period is true.