We could add `ignore_erros: yes` to the task launching the benchmark.
We could add
ignore_erros: yesto the task launching the benchmark.