-
Notifications
You must be signed in to change notification settings - Fork 92
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
benchmarks: create a ticket if zebrad mainnet sync time increases significantly #4554
Comments
At the moment, we're only running a full sync after each PR merges to the
We decided not to run a full sync on each PR, because then each PR batch would take 6 hours to merge. We could time the mandatory checkpoint sync instead, it only takes 3 hours. |
I've marked this ticket as needing a design, and being blocked by sync performance improvements. If we manage to decrease sync time by a few hours, that will change the design decisions we make. |
Mainnet zebrad + lightwalletd full sync times are currently available at: |
I rewrote the issue description instead of closing it, because we still need some way to find out when the sync is slow. |
This has been done already by automatically opening a ticket when the weekly scheduled full sync job fails in PR #7565. We also have an open ticket for doing a sync with a few tens or hundreds of thousands of blocks, and timing it: |
Motivation
Sometimes changes (like dependency updates) can affect total sync time, and we want to know this as early in the cycle as possible, so that we don't merge a performance regression.
We currently do a full mainnet sync every week. We could automatically create a ticket if the time increases significantly.
However, this still keeps the slowdown signal late in the process. But it's still better than what we have now.
Possible Implementations
This GitHub Action supports Rust (
![](https://raw.githubusercontent.com/rhysd/ss/master/github-action-benchmark/alert-comment.png)
cargo bench
) and can store and render results on thegh-pages
branch or a custom branch as a JSON blob in a.js
file, or in a separate repo, or in an arbitrary file that gets populated however you like in the workflow, say by the cache Action. Supports failing a workflow + make a comment on a PR when benchmarks fall outside a configurable threshold: https://github.com/benchmark-action/github-action-benchmark/tree/master#alert-comment-on-commit-pageSecurity
Pushing the the GH pages (if used) on external PRs may be risky: https://github.com/benchmark-action/github-action-benchmark/tree/master#run-only-on-your-branches
The text was updated successfully, but these errors were encountered: