This file keeps track of changes in the different request-log-analyzer releases. When creating a pull request, please add an entry to the “Not yet released” section in this document.
-
Nothing yet!
-
Catch SyntaxErrors when eval’ing lines.
-
Scrub parameter with file upload line in Rails requests, so it will parse successfully. (Michael Stark)
-
Allow more flexible Apache user strings
-
Rspec 3
-
Reintroduce class attribute inheritance to keep writing your own custom analyser based on an existing one simple.
-
Rubucop cleanup
-
Adapt Delayed Job 4 Log format changes (Michael Stark)
-
Catch all errors when using eval (temporary workaround to prevent crashing)
-
Fix Rubinus in Travis config.
-
Enable the sender name and email address to be specified on the command line (Joshua Pettett)
-
Dropped Ruby 1.8 support
-
Dropped ActiveRecord 2 support
-
Removed OrdereredHash
-
Removed force_encoding check.
-
Millisecond measuring units when measurements are under 1 second
-
Added specific class to html report
-
Allow floats for durations when using the %T directive in Apache format
-
Rubinius Fixes
-
Apache vhost combined now supported
-
Removed old blog from readme
-
S3 format parser more permissive, as it rejected real log files
-
Redid Travis build environment & requirements
-
Added MIT license to gemspec
-
Switched to new NGINX apache default (combined)
-
Enable database testing on Travis
-
Use bundler gem tasks for gemification
-
Rails 3 routing errors are now reported on.
-
Delayed Job 3 format now supported
-
Added NGINX support
-
Analyse & Report Rails3 partials
-
Added directories parsing support
-
Filename is now displayed in the header of the output
-
Added orderedhash class for Rubinius support
-
Fixed a number of Travis issues
-
Fixed version number being off.
-
Rake tasks are now Rails 2 & 3 compatible.
-
Ftools is replaced with fileutils.
-
Cleaned up internal API for nices Ruby integration.
-
Fixed small typo.
-
Replaced relative path reference to database_console_init with standard reference
-
Removed Encoding.default_external as it was unnecessary and caused problems with Ruby 1.9.3-head.
-
Database inserter now working for ActiveRecord 3: NOTE! The minimal ActiveRecord version required is 3!
-
Travis CI build support (travis-ci.org/#!/wvanbergen/request-log-analyzer)
-
Standard error and out are now cleaned up (#107)
-
Loading the database console now works again.
Important: The @–gets-memory-protection@ command line switch is gone. Ruby 1.9 offers this functionality without a performance penalty. If you want to keep using this option in Ruby 1.8, please do not upgrade to this version or copy paste the mixin.
Also, removed some half-finished features that were not in a usable state and are not going to be further developed, like using a database as a request source and the “FancyHTML” output format.
-
The HAProxy file format now supports more old versions.
-
The Rails 3 format now reads “parameters” lines.
-
Better parsing of hostnames and IP addresses, fixing some edge cases in all file formats that use these.
-
Change @autoload@ magic to use regular @require@ to make the source code more understandable.
-
File format specs are refactored to be more readable.
PostgreSQL format bug: we found that for some log files, about 50% of all queries will not get parsed. We are working on a fix for the file format; in the mean time you can still use it to get a feel of the performance.
-
Small bugfix when using custom file formats in Ruby 1.9.2.
-
Small fix for Rails 3 processing line edge case.
-
Set up CI testing framework. Request-log-analyzer is now tested on Ruby 1.8.7, 1.9.2 and Ruby enterprise edition.
-
Add support for the W3C log format. @–format w3c@
-
Add support for HAProxy logs. @–format haproxy@
-
Add support for providing the unit used for the Apache %D directive. Default: @%D@, or @%{micro}D@, in which case the values will be interpreted as microseconds. @%{milli}D@ will interpret them as milliseconds instead. (@%T@ interprets the value as seconds.)
-
Fixed parsing of Rails 3 Completed lines when another duration is specified (ex: Sphinx)
-
PostgreSQL format fixes
-
Fix in loading custom output formats from the command line.
-
Added parsing of object instantiation counts to Oink file format.
-
Fixed issue in Delayed::Job log parser when using multiple workers.
-
Added some stuff to the YAML result export.
-
Fixed a bug in parsing rails3’s started line for Ruby 1.9.2 Time#to_s
-
Added benchmarks of different rubies
-
Small improvements to the Rails 3 file format.
-
Fixed a issue with the –select command line option not being parsed correctly.
-
Small fixes in both the Rails 2 and Rails 3 file formats.
-
Fixed a small date handling issue in Rails 3 file format.
-
Fixed an issue with the Rails 2 file format when ActiveRecord is disabled.
-
Fixed format auto-detection issue that caused a Rails log file to show up as an Oink file.
-
Added 90-percentile interval to numerical reports.
-
Some fixes for the Rails 3 file format
-
Ruby 1.9.2 and Rails 3 compatibility fixes
-
Rails 3 log format improvements
-
Added log splitting on PID to Rails rake task @rake rla:split@ (Issue 56).
-
Several smaller bugfixes
-
Added Oink support for Rails.
-
Added @–mailsubject@ option (Issue 51).
-
Fixed require exception on empty rails folder
-
@–silent@ option implemented. No output is given when this option is used. Ideal for cron job usage (Issue 46).
-
@–gets-memory-protection@ Includes a mixin to prevent Ruby gets from using huge amounts of memory when extremely long lines are encountered in a log (Issue 47). Please note that this option will suffer a 10% performance hit.
-
Fixed a bug that disallowed line definition names that are also methods names of the Object class.
-
Initial support for Rails 3 logs. Add @–format rails3@ to your command line if auto-detection fails. <br />
Note: The log format has changed since the release of the Rails 3 beta. Only the HEAD version of Rails 3 is supported by request-log-analyzer.
-
Fixed a bug in the [[Amazon S3 access log]] parser.
-
Improved the graphs when using the @FancyHTML@ output.
-
Added a nicer API to define capture values in a file format.
-
Rails 1.9 compatibility fixes related to string encodings.
-
Fixed a bug in the request database schema creation.
-
Minor bugfixes.
- [PostgreSQL query log]
-
support. (Add @–format postgresql@ to your command if auto-detection fails.) <br />
We only had a limited sample of PostgreSQL log to implement the file format.Please support us and send in a fragment of your PostgreSQL query log file!
-
Fixed Rails action cache hit detection for newer Rails versions.
-
Added support for [[DelayedJob log]] files. (Add @–format delayed_job@ to your command if auto-detection fails.)
-
Fixed Rails 3 ActiveRecord deprecation notice.
-
Fixed a small bug in the Rails development file format report.
-
Fixed the @:multiple => true@ option for frequency trackers
-
Added a beginning for a new output format: FancyHTML, which includes charts using the Google charts API. Add @–output FancyHTML@ to your command line to activate it.
-
Improved Sinatra/Rack format auto-detection.
-
Extracted regular expressions for common log fragments, like IP addresses and timestamps, to a shared module.
-
Support parsing IPv6 IP address as well. Thanks to “dissemble”:github.com/dissemble for contributing the patch.
-
JRuby related fixes
-
Fixed a crasher bug in the HTML output of traffic reports.
-
Added file format autodetection: in most cases, providing the @–format@ parameter is no longer needed to select the correct file format. However, it is still possible to explicitly set the format if autodetection fails.
-
Added a [[MySQL slow query log]] parser (@–format mysql@). <br />
Thanks “Bart Zonneveld”:github.com/bartzon for contributing this functionality.
-
Mail host now configurable. Use @–mailhost <server>@ to specify this
-
Added specs to test mailing functionality using a mock SMTP server (mailtrap)
-
Added correct content-type to emails sent via the @–mail@ option.
-
Fixed @–mail@ command line option which was broken since version 1.4.0.
-
Added Scout & Munin plugin integration tests
-
Initial version of request-log-analyzer API.
-
Added standard deviation to duration and traffic reports.
-
Added @–report-amount <n>@ to define the number of entries in the table in the report. Use @all@ to specify that all entries should be included. The default value is 20
-
Added @–report-sort <sortings>@ options to specify how the resulting duration and traffic tables should be sorted. You can specify multiple sort fields using a comma separated list, which results in multiple tables. Supported sortings are @sum@, @mean@, @stddev@ and @hits@. The default value is @sum,mean@.
-
Restructured the [[Rails request log]] parser to be more flexible about what lines to parse, using the @–rails-format <lines>@ command line switch.
** @–rails-format minimal@ will only parse processing and completed lines. ** @–rails-format production@ will parse exceptions and cache hits besides the minimal lines. ** @–rails-format development@ will also parse SQL queries and template renderings ** @–rails-format all@ will parse every known line type. ** @–rails-format minimal,failure@ will parse the minimal lines and exceptions. Use a comma-separated list like this. to define your own line set.
-
Significantly improved the test coverage of the Rails log format.
-
Fixed a crasher bug when parsing compressed files.
-
Speed optimizations by enabling teaser checks for lines.
-
Speed optimizations to improve parsing performance of frequently called methods, i.e. @Tracker#should_update@, @Request#convert_value@, @Request#convert_timestamp@, @Request#convert_traffic@ and @Request#convert_duration@.
-
Tweaked all file formats for increased performance
-
Cleaned up and simplified some old code.
-
Added statistics about data traffic for the file formats that contain this information (Apache, Amazon S3, Rack)
-
Added support for Rack CommonLogger logs (@–format rack@).
-
Added “agent” and “referer” Apache log format constants.
-
Improved request duration parsing (%D directive) from [[Apache access log]] files.
-
Fixed a bug in the database inserter that caused tables to be created with wrong fields.
-
Added [[Amazon S3 access log]] parser. Add @–format amazon_s3@ to you command line to use it. S3 access logs are disabled by default. See “the S3 documentation on server logging”:docs.amazonwebservices.com/AmazonS3/latest/index.html?ServerLogs.html for more information to enable it.
_Retracted release_
-
Tweaking to increase performance significantly.
-
Added a @–no-progress@ command line switch to disable the progress bar parsing for an additional speed improvement.
-
E-mail option to automatically send you reports: add
--mail [email protected]to your command line. -
Allow other database connections than SQLite3 by passing a connection string to the @–database@ command line option. Examples:
** SQLite3 file database: @–database filename.db@ (as before) ** Connection URI: --database postgresql://rla:rla@localhost/requests ** ODBC-style: @–database adapter=mysql;username=root;host=localhost;database=requests@ ** Memory database: @–database adapter=sqlite3;database=:memory:@
-
Keep the previous database by default, pass the @–reset-database@ option to clear the database before parsing.
-
Added an IRB console to inspect a request database just like Rails’s @script/console@. Run with <br />@request-log-analyzer console -d <connection string>@
-
Fixed some require paths that were causing problems.
_Retracted release_
-
Improved [[Apache access log]] format. Can now handle any kind of “Apache access log file format”:httpd.apache.org/docs/1.3/mod/mod_log_config.html by providing the @–apache-format <format>@ command line argument.
-
Allow @FileFormats@ to be constructed dynamically (used for the Apache format implementation).
-
Refactored parameter passing during parsing to allow for more flexibility.
_Retracted release_
-
Added [[Apache access log]] parser. To use, add @–format apache@ to your command line invocation
-
Make the release script send the gem to RubyForge as well.
-
New, more robust gem release script using the ruby-git gem.
-
Updated outdated information in the gemspec file
-
Improved Merb log format parser
-
More robust handling of database connections in database inserter.
-
Refactored database functions
-
Cleaned up specs and tried some heckle
-
Interrupts are handled more gracefully
-
Moved munin tracker to a “separate Github project”:github.com/barttenbrinke/munin-plugins-rails/
_Retracted release_
-
Compressed logfile support
-
Parsable YAML results as output (use @–dump <filename>@ command line option )
-
“Full Rdoc documentation”:wvanbergen.github.com/request-log-analyzer
-
Introduced [[Changelog]]
-
Updated rails rake tasks
-
Ruby 1.9 support
-
Rcov coverage (use @rake spec:rcov@ to generate the coverage report)