StreamSlotMetricsRequest
StreamSlotMetricsRequest subscribes to per-slot metrics updates.
One message is published per slot terminator. That is normally an executed block, but a persistence consumer connecting without a durable marker also receives an account-bootstrap terminator, which reports no execution state: its counters are zero and it carries no state commitments.
Package: thru.services.v1
Fields
Section titled “Fields”| Field | Type | # | Description |
|---|---|---|---|
start_slot | uint64 | 1 · optional | Optional lower bound on the slots delivered. This is a FILTER on live publications, not a replay: the server holds no history and will not backfill slots that are already in the past. Slots below this value are skipped; if not specified, delivery starts from the next terminator. To read past slots use QueryService.GetSlotMetrics / ListSlotMetrics — but note those carry the counters and timestamp only, NOT the state commitments below. For the commitments the corresponding queries are QueryService.GetStateRoots and QueryService.GetActiveStateHashes, and they answer only for slots whose rows were actually persisted — a slot the indexer never committed, or one whose terminator carried no commitment, has nothing to return, so a gap is not guaranteed to be fillable. |
include_compressed_state_root | bool | 2 · optional | Opt in to compressed_state_root on the response. Off by default: the commitment costs ~34 bytes per slot and most consumers do not read it. |
include_active_state_hash | bool | 3 · optional | Opt in to active_state_hash on the response. Off by default, same reasoning as include_compressed_state_root. |