monad-ops dashboard — monad-testnet

monad-ops · monad-testnet

What this is. Public operational dashboard for a Monad testnet validator node operated by rustemar (validator alias hukutu4). Live-updating — shows retry rates, throughput, chain integrity, and alert history. Open-source (repo).

last block
last seen —
blocks / sec (1min)
— tx/s
retry_pct avg (1min)
5min avg —
retry_pct peak (1min)
share of tx re-executed after parallel conflict
current epoch
epoch info loading…
retry_pct ✓ <25 ○ <65 △ <75 ⚠ ≥75
blocks/s ✓ ≥2.0 ○ ≥1.5 △ ≥0.5 ⚠ <0.5
last block age ✓ <3s △ <10s ⚠ ≥10s
validator timeouts ✓ <1% ○ <3% △ <5% ⚠ ≥5%
local pacemaker fires/min ✓ <1 ○ <3 △ <6 ⚠ ≥6
thresholds set from historic retry_pct distribution (worst case 92%, 2026-04-20 stress batch); validator-timeout target from the 2026-04-20 stress-test summary
effective tps peak (1min)
avg — tx/s per block
gas/sec effective peak (1min)
peak gas/sec inside a single block
validator timeouts % (5min, chain-wide)
Foundation target <3% · share of rounds closed by TimeoutCertificate
local pacemaker fires (5min, this node)
average per minute · operator-side complement to chain-wide TC %
tip lag vs reference RPC
local
vs
reference
delta
compared against testnet-rpc.monad.xyz · answers "is my node lagging, or is the whole network halted?"
Peak = the fastest moment inside a single block (not a sustained rate). Testnet blocks arrive slowly (~2.5 blk/s), so the average stays low while per-block peaks reveal Monad's parallel-execution headroom.
period
retry_pct retry_pct — share of tx re-run per block
tx/block tx count per block
validator timeouts % share of consensus rounds closed by TimeoutCertificate · per-minute
base fee (gwei) per-block base_fee from monad-bft proposals · downsampled per window
network-layer signal events monad-bft RaptorCast / wireauth / consensus-state signals · per-minute · stacked
pre-finalization divergence events single divergences vs cluster-tier · per-minute · stacked
block execution time breakdown time per execution step, in microseconds · stacked
state_reset tx_exec commit
execution strain composite ratio across retry_pct · tx_exec · commit · state_reset — 1.0 = baseline · per-bin
1.0 baseline · >1.3 elevated · >1.5 high
host probes periodic local health checks · services · disk · open-file limits
  • probes running…
node version installed monad package vs. apt repo · hourly
checking apt repo…
recovery-path stall watch runloop_monad.cpp recovery-assertion class · public archive
loading incident archive…
active validator set staking precompile · refreshed every 5 min · changes per epoch
querying staking precompile…
chain integrity reorg detector · invariant check
reorgs observed since process start
The counter reflects pre-finalization block-id divergence at the execution layer — observed when the same block number arrives with a different exec-layer id. HotStuff-2 expected behaviour at depth 0–1; the canonical chain finalizes correctly regardless. Single events are informational; the card tints amber when ≥3 events land in the last 24h (cluster threshold).
top contracts by re-execution rate ranked by re-execution rate · —
contract blocks retried retried ratio avg rtp tx
collecting contract labels…
critical incidents assertions · panics · chunk-exhaustion events · critical host-probe failures
  • ● clear — no critical events in buffer
recent alerts 0 active · full history →
  • no alerts in buffer