Overview

Token, traffic, reliability and latency analytics

Token usage over time

Top API keys

Usage by hour of day

Reliability

Hourly token usage

HourRequestsInputOutputTotalCachedErrors

24-hour profile

Hour of dayRequestsInput tokensOutput tokensTotal tokensErrors

API key × hour token matrix

Total tokens grouped by local hour. Useful for identifying which key drives peak load.

Coding-agent sessions

SessionAPI keyRequestsInputOutputTotalAvg latency

Reasoning efficiency

Effective gateway effort after API-key policy and compatibility normalization.
EffortRequestsTotal tokensReasoning tokensCached tokensAvg latency

Create API key

API key policies

NameScopePrefixStatusBackendRPMTPMConcurrencyMax inputMax outputContextReasoningMaxLast used

Tenants

NameStatusConcurrency

Projects

TenantProjectStatusConcurrency

Virtual models / Thinking profiles

AliasUpstreamStatusThinkingEffortDescription

Usage by tenant

TenantRequestsTokensErrors

Usage by project

ProjectRequestsTokensErrors

Audit trail

TimeActorActionResourceID

Usage by model

ModelRequestsTokensErrorsError rate

Recent requests

TimeAPI keyModelEndpointStatusInputOutputTotalLatencyTTFTSessionError