Playground

Playground

Trace

/api/v1/telemetry/traces

A trace/span record.

listget

Schema

FieldTypeNotes
id*uuidServer-assigned ULID with type prefix (e.g. per_…).
tenant_id*stringTenant scope — auto-stamped from the caller's JWT.
app_idstringApp scope. Stamped ONLY when the caller's JWT was minted for a specific Application (integration API keys). Absent for human-user sessions. Filters reads when present.
created_at*date-timeServer stamp.
updated_at*date-timeServer stamp; updated on every patch.
created_byuuidPerson id from the caller's JWT (sub).
updated_byuuidPerson id from the last writer's JWT (sub).
sourcestringProvenance tag — defaults to 'edm'.
source_typeenumfrontend | backend | server | system | apiWhere the write originated. Defaults to 'api'.
is_deletedbooleanSoft-delete flag. Excluded from default list queries.
deleted_atdate-timeStamped when soft-deleted; null otherwise.
deleted_byuuidPerson id who soft-deleted; null otherwise.
schema_versionnumberDocument schema version. Bumped on incompatible writes.
parent_span_idstringDistributed-tracing parent span ID, if this trace is a child span.
agent_run_idstringIdentifier of the agent run this trace belongs to, if part of a multi-step agent workflow.
user_idstringIdentifier of the user who initiated the LLM call. Null for system-initiated calls.
api_key_idstringIdentifier of the virtual API key used for this call.
modelstringModel identifier in 'provider/model-name' format.
providerstringLLM provider name.
span_kindstringKind of LLM operation represented by this trace.
input_tokensnumberNumber of prompt/input tokens consumed.
output_tokensnumberNumber of completion/output tokens generated.
cache_read_tokensnumberNumber of tokens served from the prompt cache.
cache_write_tokensnumberNumber of tokens written into the prompt cache.
latency_msnumberEnd-to-end latency of the LLM call in milliseconds.
cache_hitbooleanWhether this response was served (fully or partially) from the prompt cache.
cost_usdnumberEstimated cost of this LLM call in USD.
finish_reasonstringReason the model stopped generating. One of 'stop', 'length', 'tool_calls', or 'error'.
routing_rulestringIdentifier of the routing rule that selected the provider for this call.
policy_decisionsjsonPolicy decisions evaluated for this call by the coreiq policy engine.
security_violationsjsonPII or DLP violations detected by the policy engine for this call.
request_hashstringSHA-256 hash of the canonicalized request payload, used as the cache key.
episode_idstringLinked evolution episode identifier, if this trace is part of a tracked episode.
tsdate-timeISO 8601 UTC timestamp of when the LLM call occurred.

API

Loading manifest…