docs: add 2026-07-15 multi-chain capacity report - #317
Open
tvinagre wants to merge 1 commit into
Open
Conversation
Per-chain capacity for base, arbitrum, polygon, unichain, and bsc from the staging measurement campaign, plus the ethereum synthetic-vs-real calibration. Includes the recomputed strict/knee/wall table, extrapolated c6a and tier-2 estimates, launch sizing with the reused cost model, and operational findings. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds
docs/scaling/2026-07-15-multichain-capacity.mdand the raw ladder JSONsunder
docs/scaling/data/2026-07-15/.The report covers the staging capacity campaign for base, arbitrum, polygon,
unichain, and bsc, plus an ethereum synthetic-vs-real calibration. It contains:
latencies, unsolved rate, strict/knee/wall per tier). Every number was
recomputed from the per-step JSON data, not copied from the run summaries.
generate-requestssets match the real 10kaggregator set on ethereum with a 0% capacity delta, with the caveat that
unsolved rates do not transfer.
cadence sets the latency tail (arbitrum's 0.25s blocks give the fattest idle
tail); the instance lottery (~2x t3a vs c6a) reconfirmed; unichain never
failed on one worker; and 1-worker to 2-worker vertical scaling.
ESTIMATED (the c6a reruns were blocked by an AWS fleet-quota outage).
startup-probe overruns, generator memory, and the fleet-quota outage).
Ethereum's production numbers stay authoritative from the 2026-07-13 report
(PR #307). t3a/flex rows are a floor pending c6a mop-up reruns.
🤖 Generated with Claude Code