Response signals and errors
This is the reference for what Lasso Cloud adds to responses and the errors it returns. Every error Lasso originates names its fix innext_action. For how
to use these signals together, read routing evidence.
Response headers
lasso_meta
Add ?include_meta=body for a top-level lasso_meta object, or
?include_meta=headers for the x-lasso-meta header. Lasso Cloud returns:
Read serving access and fallback from the
x-lasso-access and
x-lasso-fallback headers. Managed providers appear under public aliases;
Custom profiles show your provider names. The routing engine’s full field list, including block
protection fields, is in request metadata.
WebSocket callers add "lasso_meta": "notify" to a request for a follow-up
lasso_meta notification.
RPC errors
RPC errors stay standard JSON-RPC. Errors Lasso originates carryerror.data.code, error.data.next_action and error.data.retryable. A
retryable error with Retry-After is safe to repeat as is after the wait.
These are the requests Lasso answers itself, before routing:
Errors from providers pass through unchanged. When every eligible provider
fails, the final JSON-RPC error arrives in an HTTP 200 response. An
eth_getLogs range a provider refuses as too large returns -32005 with
data.code: "log_range_limit"; split the range and retry.
WebSocket close reasons
Revoking a key, or the end of a rotation overlap, closes sockets using that
secret.
Management API errors
Management errors use one envelope:
Exact response schemas are in
openapi.json.
Next
- Routing evidence: use these signals to debug.
- Upgrading from Agent API v2: older headers and error fields.