[dogstatsd] handle properly utf8 packets #1279
Merged
Conversation
This file contains 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
LeoCavaille
force-pushed
the
leo/dogstatsdutf8
branch
from
e5d0121 to
1d35512
Compare
| # network socket, but if submit_packets is used | ||
| # programatically and packets is unicode already | ||
| # then do not decode! | ||
| if not isinstance(packets, unicode): |
There was a problem hiding this comment.
Could you use:
if not type(packets) == unicode:instead ?
It's slightly faster:
topdog@i-1b871af7(prod):~$ python -m timeit 'not isinstance("sdfsdfsd" , unicode)'
10000000 loops, best of 3: 0.197 usec per looptopdog@i-1b871af7(prod):~$ python -m timeit 'not type("sdfsdfsd") == unicode'
10000000 loops, best of 3: 0.151 usec per loopand this code is run in a pretty performance sensitive loop.
|
Looks great besides the comment! |
|
what's the perf impact? |
|
Looks like the extra Benchmark utf8 handling (only the code in this branch, raises on the master branch) Benchmark ascii text with this branch Benchmark ascii text with master branch |
|
Also @remh I tried your suggestion but looks like it results in slower code on my laptop: |
|
30% is a lot. Does the statsd protocol mention anything about accepting UTF-8 in the payloads ? |
|
@remh Using The performance impact is only 7.1 % now. About the statsd protocol, didn't find much in the protocol, however I found a few |
LeoCavaille
force-pushed
the
leo/dogstatsdutf8
branch
from
18e2104 to
a7ab7fc
Compare
|
So 7% is still pretty high i think. It might be worth making that optional ? @clutchski any additional thought ? |
LeoCavaille
force-pushed
the
leo/dogstatsdutf8
branch
from
a7ab7fc to
0ee4f16
Compare
LeoCavaille
force-pushed
the
leo/dogstatsdutf8
branch
from
0ee4f16 to
d610cc4
Compare
|
@remh this one is ready to go too, however I merged the histogram min branch in it before because it contained a refactor of aggregator.py, thus that one should go out first. |
Fixes #1256. We should always consider that dogstatsd receives a utf-8 encoded string through its socket, though it comes with a non-negligeable performance overhead over ~7% in simple benchmarks. The default behavior of the server is thus considering packets as ASCII-only content, flipping the utf8_decoding flag in the config will allow you to parse such packets correctly. Most (dog)statsd clients should already encode their packets in utf-8 when sending data.
LeoCavaille
force-pushed
the
leo/dogstatsdutf8
branch
from
d610cc4 to
ebf3625
Compare
LeoCavaille
added a commit
that referenced
this pull request
Jan 28, 2015
[dogstatsd] handle properly utf8 packets
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Fixes #1256. We should always consider that dogstatsd
receives a utf-8 encoded string through its socket,
but still support unicode python strings in case we
submit things programatically (e.g. useful for tests)