Crouton's status is measured by synthetic probes running from our eu-central region every 30 seconds. Each probe issues canonical health-check methods (eth_blockNumber, eth_chainId for EVM; status for Cosmos) against each chain endpoint.

Current status (the green/yellow/red badge) is classified per 1-minute bucket: a minute is green when probe success rate ≥ 95%, p95 latency stays within the chain's SLO threshold, and block lag stays within 3 blocks for EVM chains and 5 blocks for Cosmos chains.

Long-term uptime (30d/90d percentages) is calculated from hourly probe aggregates. Block lag is checked at the 1-minute granularity for current status only — hourly buckets do not include block-lag data. Uptime figures therefore reflect success rate and p95 latency across the reported window.

The denominator for each percentage is the number of hours in which the chain was actually probed, not a fixed 30×24 or 90×24. Recently-added chains, or chains with brief data gaps, will show an asterisk (*) next to their uptime number; hovering exposes how many hours of probe data the percentage was computed over. We surface the actual window so a chain with two days of history doesn't silently appear identical to a chain with thirty.

Latency measurements are end-to-end including TLS termination but excluding DNS resolution. We publish all incidents with at least 5 minutes of measurable impact in our probes, including those affecting only specific methods or chain regions.

→ Read full methodology → See our open-source probe configs → Subscribe to incident updates