ci: make span start max_rss SLO consistent across scenarios

[brettlangdon]

Description

We are seeing some wider ranges than these SLOs, so updating the max_rss to be consistent across all the SLOs. Seems some of them need a slightly higher limit anyway.

Testing

Risks

Additional Notes

[datadog-datadog-prod-us1]

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🔄 Datadog retried 1 test - 1 passed on retry

🚧 96 tests that failed were ignored due to quarantine

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 008af0c | Docs | Datadog PR Page | Give us feedback!}

[pr-commenter]

Benchmarks

Benchmark execution time: 2026-05-15 14:40:52

Comparing candidate commit 008af0c in PR branch brettlangdon/bump.max_rss.lso with baseline commit 1ab9c6f in branch main.

Found 0 performance improvements and 4 performance regressions! Performance is the same for 593 metrics, 10 unstable metrics.

scenario:iastaspects-stringio_aspect

• 🟥 execution_time [+582.479µs; +629.633µs] or [+14.961%; +16.173%]

scenario:iastaspectsospath-ospathbasename_aspect

• 🟥 execution_time [+98.029µs; +104.155µs] or [+22.489%; +23.894%]

scenario:span-start

• 🟥 execution_time [+1.499ms; +1.674ms] or [+9.661%; +10.789%]

scenario:telemetryaddmetric-1-count-metric-1-times

• 🟥 execution_time [+167.217ns; +195.143ns] or [+7.952%; +9.280%]

[cit-pr-commenter-54b7da]

Codeowners resolved as

.gitlab/benchmarks/bp-runner.microbenchmarks.fail-on-breach.template.yml  @DataDog/python-guild @DataDog/apm-core-python