sander/perfix
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update README.md

Browse source
This commit is contained in:
Sander Hautvast 2018-05-19 00:50:48 +02:00 • committed by GitHub
parent 678b122c33
commit 32a64ced1c
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
README.md
View file

@ -3,9 +3,9 @@ Pretty basic profiling tool for JVM's
# Highlights:
* Meant for development time (after process stops, data is gone).
* Minimal memory footprint (agent is 2.5 mb).
* Minimalistic commandline interface.
* Execution time is measured in nanoseconds, reported in milliseconds (this way the totals and averages are most precise, but also human readable).
* Minimal memory footprint (agent < 1 mb).
* Minimalistic web interface.
* Execution time is measured in nanoseconds
* No manual instrumentation necessary using loadtime bytecode manipulation (javassist).
* No special jdbc configuration necessary (ie no wrapped jdbc driver).
* The agent is also the server (unlike commercial tooling). This way there is no overhead in interprocess communication.
@ -13,19 +13,16 @@ Pretty basic profiling tool for JVM's
# Usage
* Agent that instruments loaded classes: -javaagent:<path>/perfix.jar
* Include classes for instrumentation with -Dperfix.includes=com.project. ...etc (includes subpackages)
* Ssh interface to report executed methods and sql query excutions:
* Web page to report executed methods and sql query excutions:
<br/> * #invocations
<br/> * total execution time for the method in nanoseconds (this is also the sorting order)
<br/> * average time in nanoseconds per method (= total/#invocations)
* The (ssh) server starts on port 2048 by default. Use -Dperfix.port=... to adjust.
* The http server starts on port 2048 by default. Use -Dperfix.port=... to adjust.
# roadmap
* Overhead (in method execution time) not clear yet. I wouldn't use it in production.
* Finish jdbc query logging
* Make output format configurable
* Implement password login (now any)
* Add web interface (maybe)
* Finish jdbc query logging (CallableStatement)
* Implement an actual call stack the way commercial tools work
# DISCLAIMER: