on-chain observability · reactive network · lasna testnet
62% of the network's subscriptions are dead — nobody notices.
Reactor Watch listens to your reactive contracts, catches silence, reverts and dead
configs — and drops an alert into a public on-chain mailbox, readable forever with a plain
eth_getLogs. It notices silence. It does not touch your funds.
100% on-chain · no servers · no API keys · vanilla static page
[ live ] · watcher v5.6
network incident: callback delivery paused callback delivery: livestand filters: 12/12 Active · fed by scan passes every ~1000 blocks
live paused — showing last known values
[ how it works ]
Three steps, zero trust required
- We subscribe to your events — DeployerSubscribe, topics, config lifecycle.
- The Reactive VM catches silence, REVERTS, DEATH, low reserve — between blocks.
- An alert lands in the public mailbox — on-chain, permanent, plain eth_getLogs.
[ what it catches ]
Eight ways a subscription dies quietly
SILENCE
Your contract stopped emitting. The watcher notices the gap and alerts.
REVERTS
Callbacks keep failing. Each revert is caught and reported.
DEAD
Node blacklist or paused config — the death event does not go unnoticed.
REVIVED
Back from the dead — you learn it woke up, and when.
LIFECYCLE
Registrations, unregisters, config churn — the full lifecycle of your contracts.
LOW_RESERVE
Debt kills delivery quietly. Watch the reserve before the node does.
EXPIRED
Slots that expired from quiet — archived, with a trace left behind.
HEARTBEAT
The watcher itself is watched: a self-check pulse, on-chain.
[ the numbers ]
Dead subscriptions are the norm, not the exception
[ the whale ]
One operator: 5,230,000 transactions, 535 dead subscriptions, 7 months of silence.
nobody noticed for months — that is the product
snapshot 2026-09-08 · recalculated hourly by watcher-cron
[ why now ]
Two clocks are ticking
Chainlink Automation is shutting down
Every automation user must migrate somewhere. Migrations break subscriptions — and broken subscriptions stay broken unless something notices. We are that something.
Callbacks on Ethereum start failing
The second gas meter lands, and reactive callbacks that write new state begin to fail. The foundation itself named the class of breakage — Reactor Watch is the protocol-native detector for exactly these failures.
[ for the agents ]
The network's own CTO calls subscriptions the sense organs of on-chain agents. 62% of those organs are dead — and nobody notices. Reactor Watch is the pain reflex of the nervous system: observability for every subscriber, agents included.
[ pricing ]
Usage-based, on-chain, returnable
Cost plus 15%. You pay per what the watcher actually does — nothing else. The deposit is yours and withdrawable on-chain.
| unit | price |
|---|---|
| event delivered to your watcher | 3 milli lREACT |
| alert posted | 30 milli lREACT |
| scan pass | 1 milli lREACT |
| deposit (returnable) | from 3 lREACT |
withdraw() is live in v5.6 · two-phase refund, claimable on-chain
[ your stand ]
Register from your wallet — that's the whole onboarding
demo loads below…
register() on Watcher v5.6 · deposit is counted per milli and returnable · Lasna chainId 5318007
[ quickstart ]
cast send 0x810E…03F5D "register()" \ --value 3ether \ --rpc-url https://lasna-rpc.rnk.dev/full quickstart & cookbook — coming with the builders' guide
[ dev fund ]
For builders who are actually shipping: grant track requests for automation & observability tooling. If your product runs on reactive subscriptions, it needs a watchdog — talk to us before Glamsterdam does it for you.