GET /metrics/cycles/:n
pox5-reader coverage for one reward cycle, the Hiro cycle record and the cycle's distributions.
curl https://metacenter.0xo.in/api/metrics/cycles/143| Field | Source |
|---|---|
intervals_computed, pool, obligation, bond_paid, stx_paid, shortfall, coverage, headroom | get-coverage-for-cycle when a public node will run it (onchain), otherwise summed from the cycle's calculate-rewards events (mirrored). Each field says which, and names the distributions it summed. |
hiro | Hiro /extended/v3/staking/cycles/:n (mirrored). Note: rewards.btc.waterfall.bonds there is the amount paid, not owed. |
intervals | The cycle's distributions, as in /intervals |
Returns 400 when n is not a non-negative integer.
Which source answered
get-coverage-for-cycle reads pox-5 several times, so a public node refuses it for every cycle (see Read limits). Rather than answer with a column of nulls, this route falls back to the cycle's own calculate-rewards events — the same events the contract's figures are reconstructed from — and labels the result mirrored, naming the distributions it summed:
{
"value": 17.858600535168804,
"unit": "x",
"provenance": "mirrored",
"source": "calculate-rewards events for distributions 286–287",
"note": "Summed from this cycle's calculate-rewards events, not computed by pox5-reader: pox5-reader is deployed on mainnet, but this read-only cannot be called through the public Hiro API: it calls into pox-5 four or more times, each about 136k of read length, over the 500,000 cap. It is checked against live mainnet state by contracts/scripts/verify-at-tip.mjs. The dashboard's headline coverage is the contract's own figure for the current cycle so far, so the two cover different windows and can differ by a couple of sats per interval from integer truncation."
}Showing coverage; the full response has more fields.
Why this can differ from the dashboard
Both are right; they answer different questions.
| This route | The dashboard headline | |
|---|---|---|
| Window | One whole reward cycle: every distribution booked to it, summed | The current cycle, realised so far |
| Source | The cycle's calculate-rewards events | pox5-reader, through coverage-cache |
| Label | mirrored | onchain |
So a finished cycle and the cycle in progress will not match, and should not. Even for the same window, the contract rebuilds the pool from rewards-per-token × shares, which can land within a couple of sats of the event total through integer truncation; npm run verify allows exactly that tolerance and no more.
Live response (excerpt)
{
"btc": {
"total": "230327835",
"claimed": "173670885",
"waterfall": {
"bonds": "13810222",
"stx_only": "184039972",
"reserve_deposit": "32477641"
}
}
}Showing hiro.value.rewards; the full response has more fields.