perf: human log sink performance bump via reduction of allocs/total allocated memory #220
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Callum Styan <callumstyan@gmail.com>
Signed-off-by: Callum Styan <callumstyan@gmail.com>
…e the optimized version should return an error Signed-off-by: Callum Styan <callumstyan@gmail.com>
Pull Request Test Coverage Report for Build 3cb927f1a22a74cedce94eeec74e36f7dad4396e-PR-220Details
💛 - Coveralls |
cstyan
changed the title
Signed-off-by: Callum Styan <callumstyan@gmail.com>
Signed-off-by: Callum Styan <callumstyan@gmail.com>
…ne very specific case that should never happen if we're actually going to OOM via logging it would panic Signed-off-by: Callum Styan <callumstyan@gmail.com>
… files for new test cases Signed-off-by: Callum Styan <callumstyan@gmail.com>
Signed-off-by: Callum Styan <callumstyan@gmail.com>
Emyrk
reviewed
Nov 14, 2025
Comment on lines
-108
to
-113
| level := ent.Level.String() | ||
| level = strings.ToLower(level) | ||
| if len(level) > 4 { | ||
| level = level[:4] | ||
| } | ||
| level = "[" + level + "]" |
Member
There was a problem hiding this comment.
wow this old code was kinda dumb in comparison.
Comment on lines
+93
to
+110
| func bracketedLevel(l slog.Level) string { | ||
| switch l { | ||
| case slog.LevelDebug: | ||
| return "[debu]" | ||
| case slog.LevelInfo: | ||
| return "[info]" | ||
| case slog.LevelWarn: | ||
| return "[warn]" | ||
| case slog.LevelError: | ||
| return "[erro]" | ||
| case slog.LevelCritical: | ||
| return "[crit]" | ||
| case slog.LevelFatal: | ||
| return "[fata]" | ||
| default: | ||
| return "[unkn]" | ||
| } | ||
| } |
|
|
||
| // signed ints | ||
| case int: | ||
| var a [20]byte |
Member
There was a problem hiding this comment.
is the 20 arbitrary? Can it be less? Or can it be overflowed with very large numbers?
Strings are not handled in this switch statement. Is it because strings are of arbitrary length? Can we handle them quicker than the formatValue? I ask because I imagine most of our types logged are strings.
Note, zerolog handles this by making the typed constructors. Instead of a generic slog.F("", <any>), they do slog.Int32/String/Float64/etc("", <typed>).
| @@ -0,0 +1,241 @@ | |||
| diff --git a/internal/entryhuman/entry.go b/internal/entryhuman/entry.go | |||
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I was poking around in profiling data for our dogfood instance and I noticed that over a 24h period debug logging through the human readable log sink was producing 10-12% of our memory allocation. Garbage collection is only ~7.5% of our CPU time, but hopefully this has a nice reduction for allocs/memory allocated and small reduction for GC related CPU usage.
Benchmark data:
Note that the existing benchmark didn't cover many use cases, so the changes are not as obvious there:
Current alloc space:
Current alloc object:
Alloc objects path:
Alloc space path:
GC CPU time:
