Measurement.
8 stories
Is xOne later than geth-XDC to a new block? A same-host sample, and two tool problems
A 10-minute same-host sample of when xOneRS and geth-XDC first report each net2040 block, why measure-head-lag.sh cannot answer it, and a null latest block.
Measuring an xOne node before and after a change with measure-node.sh
How ops/bench/measure-node.sh records RPC latency, block import times, disk I/O and memory of a running xOne node before and after a window, and only reads.
A first run of measure-node.sh on a net2040 xOne node: latency, memory and I/O
We ran measure-node.sh for eight minutes on a net2040 xOne follower: RPC latency for seven calls, memory split, disk I/O, and errors the tool cannot explain.
Why net2040 makes a block every 4.5 seconds today: two late xOne seats
net2040 averaged 4.0 to 4.5 s per block on 2026-11-15, not 2 s. No round timed out: the Python and C++ xOne seats sealed 13 to 27 seconds after their parent.
reth-xdc on Apothem: a slow catch-up measured over nine samples
A reth-xdc minimal node on Apothem held a gap of about 48,900 blocks for seven hours, then gained about 30 blocks a minute. The samples, rate and ETA.
Memory per XDC client: resident memory of 33 nodes from the fleet probe
Resident memory (RSS) of 33 XDC nodes by client and sync mode, from eight read-only fleet samples, and why file-backed pages make some numbers look larger.
How fast does an XDC datadir grow? Growth per hour and per block for 30 nodes
Datadir growth of 30 XDC nodes over 8.8 hours of read-only samples: per hour and per block, on Apothem and net2040, with one outlier that is now an issue.
RPC read latency of XDC clients: a 20-call sample on 33 nodes
Median and p95 latency of eth_blockNumber, eth_getBlockByNumber and eth_getBalance on 33 XDC nodes and three hosts, from a read-only sample, and how to read it.
Edit this page on GitHub opens the page source (app/blog/tags/[tag]/page.jsx). GitHub may ask you to sign in. Report a correction opens a prefilled GitHub issue for this page.