feat(layer)!: make adding multiple dynamic fields cheaper

[mladedav]

Motivation

Adding multiple dynamic fields was expensive because it had to always allocate the string for each key on every tracing invocation as well as create a serde_json::Value.

Solution

Now the strings may be &'static str if known beforehand and the values may be left as impl Serialize if they're available. The fields will be serialized into the output string as needed.

[github-actions]

Rust Benchmark (1.89.0)

Details

Benchmark suite	Current: d5faf2c	Previous: b24ff4a	Ratio
new_span/single_thread/1	316 ns/iter (± 4)	287 ns/iter (± 11)	1.10
new_span/multithreaded/1	22674 ns/iter (± 1163)	39198 ns/iter (± 2227)	0.58
new_span/single_thread/10	3162 ns/iter (± 13)	2872 ns/iter (± 18)	1.10
new_span/multithreaded/10	27591 ns/iter (± 1180)	41613 ns/iter (± 1981)	0.66
new_span/single_thread/50	15766 ns/iter (± 303)	14399 ns/iter (± 215)	1.09
new_span/multithreaded/50	56669 ns/iter (± 2876)	56299 ns/iter (± 3008)	1.01
event/root/single_threaded/1	767 ns/iter (± 11)	835 ns/iter (± 5)	0.92
event/root/multithreaded/1	22756 ns/iter (± 751)	39424 ns/iter (± 2299)	0.58
event/unique_parent/single_threaded/1	1003 ns/iter (± 6)	1112 ns/iter (± 6)	0.90
event/unique_parent/multithreaded/1	25631 ns/iter (± 899)	44411 ns/iter (± 2604)	0.58
event/shared_parent/multithreaded/1	23573 ns/iter (± 956)	40319 ns/iter (± 3093)	0.58
event/multi-parent/multithreaded/1	28332 ns/iter (± 1159)	41106 ns/iter (± 2016)	0.69
event/root/single_threaded/10	7671 ns/iter (± 34)	8221 ns/iter (± 108)	0.93
event/root/multithreaded/10	34475 ns/iter (± 1885)	46927 ns/iter (± 2828)	0.73
event/unique_parent/single_threaded/10	10037 ns/iter (± 59)	11189 ns/iter (± 46)	0.90
event/unique_parent/multithreaded/10	44032 ns/iter (± 1980)	53694 ns/iter (± 2472)	0.82
event/shared_parent/multithreaded/10	40493 ns/iter (± 1439)	48937 ns/iter (± 2569)	0.83
event/multi-parent/multithreaded/10	66042 ns/iter (± 2744)	80030 ns/iter (± 2840)	0.83
event/root/single_threaded/50	38293 ns/iter (± 106)	41238 ns/iter (± 727)	0.93
event/root/multithreaded/50	91483 ns/iter (± 5168)	106986 ns/iter (± 7248)	0.86
event/unique_parent/single_threaded/50	50328 ns/iter (± 308)	55732 ns/iter (± 689)	0.90
event/unique_parent/multithreaded/50	129107 ns/iter (± 9187)	132678 ns/iter (± 7031)	0.97
event/shared_parent/multithreaded/50	118203 ns/iter (± 11101)	131620 ns/iter (± 7785)	0.90
event/multi-parent/multithreaded/50	416102 ns/iter (± 26399)	418921 ns/iter (± 28946)	0.99

This comment was automatically generated by workflow using github-action-benchmark.