Leo Blöcher
90d · built 2026-09-08
Performance
What Leo Blöcher shipped in the selected window, measured in ETV, and how it compares with the 90 days before it.
Effective capacity
−0.7engineers
delivers like 0.3 (0.3x pre-AI)
Output (ETV)
0.8ETV
−83.0% vs 4.8 prior
Features share
9.9%
−40.4 pp vs prior window
Fixes share
8.6%
−11.9 pp vs prior window
Work mix
9.9% Features67.9% Maintenance11.1% Tests2.5% Docs8.6% Fixes
18 commits over 90 days, ending 2026-09-08.
Daily performance
Daily ETV, stacked by Features, Maintenance, Tests, Docs and Fixes.
Repository spread
Where this developer's commits land. Concentrated work (top1 > 80%) vs polymath spread (top1 < 30%).
| Repo | Commits | ETV |
|---|---|---|
| foundations | 11 | 0.8 |
Most impactful commits
Top 10 by ETV in the last 90 days.
- 0.2ETVCache `SharedSpan::is_sampled` for deferred user tracing spans Relaxed atomic load and stores are just regular load and stores on x86 and arm64, so this adds zero cost for regular spans and allows us to avoid some more lock accesses for deferred user tracing spans.github.com-cloudflare-foundations · 31736a72 · 2026-09-03
- 0.2ETVSplit tracing metrics to allow multiple pipelines In #223 we are adding a second tracing pipeline. To avoid clashing metrics between the two, we need to add a label to distinguish them.github.com-cloudflare-foundations · 3f0d07cb · 2026-07-10
- 0.2ETVFix potential double panic when logging inside panic hook Inside a panic hook, any new panic is automatically a double panic and aborts the process. We very much would like to avoid this, but our logger can panic when it fails to do output. This means we need to work around panics from our logger. To do so, we lift the `Fuse` wrapper that converts `slog::Drain` errors into panics from our root drain into the `build_log_with_drain` helper (where we also add log level filtering.) We can then access the unfused root drain directly when logging panics and ignore errors returned from it. A new internal `log_to_drain` helper function makes this available to the entire foundations crate.github.com-cloudflare-foundations · 12199d5b · 2026-06-29
- 0.1ETVAdd regression tests for panic hook logging - Ensure context fields are included in the panic log. - Verify that an error in the `slog::Drain` does not cause the process to abort.github.com-cloudflare-foundations · fb9ce7c1 · 2026-06-29
- 0.0ETVSetup bonk workflow Adapted from workerd's `bonk.yml`.github.com-cloudflare-foundations · 1cdf7f93 · 2026-06-19
- 0.0ETVchore: Releasegithub.com-cloudflare-foundations · 88036c7c · 2026-08-13
- 0.0ETVAdd cfg(foundations_jemalloc_disable) to opt out of global jemalloc allocator Usually our advice would be to not enable the `jemalloc` cargo feature, but this becomes difficult in large dependency trees where a user may not necessarily control all dependencies that in turn depend on foundations. Providing a flag to disable the `#[global_allocator]` specifically allows such projects to build anyway. This is also useful to enable ASAN builds, which otherwise wouldn't cover allocations initiated from Rust.github.com-cloudflare-foundations · 4338e786 · 2026-06-10
- 0.0ETVfix: Make ScrapeFormat non-exhaustive to allow future expansiongithub.com-cloudflare-foundations · 3976f3bc · 2026-08-13
- 0.0ETVFix bonk workflow by downgrading opencode The workflow doesnt work. It works in other repos which pin an older opencode version. Opencode 1.17.8 included a change to authorization for CF AI Gateway. Try downgrading to the version before that to see if it fixes the workflow. Also switches the model to Opus 4.8.github.com-cloudflare-foundations · 4f4f7c7a · 2026-06-22
- 0.0ETVMove cfg(foundations_generic_telemetry_wrapper) docs out of lib.rsgithub.com-cloudflare-foundations · 9c16d397 · 2026-06-10