Haproxystats Versions Save

This is a bug fix release.

Changelog:

• Fix conditional timing metrics generation
• Use the network target on the Systemd unit files
• Calculate average only when actual connections have been made
• Update binary path in systemd file for haproxystats-process service
• Update binary path in systemd file for haproxystats-pull service
• Update python packages for CentOS distribution

This release adds a thread to check if the workers of haproxystats-process* are alive. This thread will exit the main program if at least one worker isn't alive. Last but not least, @crovner fixed a runtime error, thanks dude.

Changelog:

• Fix runtime error when specifying a custom list of metrics
• Check liveness of workers and exit if at least one is dead
• Fix formatting in README for group-namespace option

This release adds only one new functionality about grouping metrics for frontend, backend and server names based on patterns. Please read group-namespace option of graphite section for more details.

Changelog:

• Support grouping metrics by frontend, backend and server names

This is a minor release which fixes an issue introduced with the 0.4.1 version.

Changelog:

• Add missing comma

This is a minor release which updates Ansible playbook and adds support for processing total requests for backends.

Changelog:

• RELEASE 0.4.1 version
• Add supporting of ubuntu hosts and update Ansible playbook
• Also report about total requests for backend
• update readme file
• update Ansible Playbook and readme file

This release adds support for pulling statistics from TCP sockets, which allows users to pull statistics from remote HAProxy servers. There are some minor changes in the code in terms of readability and PEP257 compatible docstrings. We also fix a crash when we process statistics from HAProxy version 1.5 as "CurrSslConns" metric doesn't exist in that version. Last but not least, this release adds support for pulling few more statistics (see 4608202d10).

Changelog:

• RELEASE 0.4.0 version
• Handle the absent of daemon's metric
• It's a counter, so should be in different section
• Update the list of supported metrics
• Use int rather string in the default values
• Use int rather string in default value
• Switch to correct section name for pycodestyle
• PEP257 compatible docstrings
• Switch from yarl.UR to urlparse from standard Python library
• Update reference of UNIX sockets to include TCP
• PEP257 compatible docstring
• Simplify the return logic of two functions
• Add configuration_check_for_servers()
• PEP257 compatible docstrings
• Disable pylint warning for too-many-branches
• Change the log severity to info for task report
• Add support for pulling statistics from TCP sockets
• Make that code more readable
• PEP257 compatible docstrinigs
• Fix typo in README

This release brings minor improvements in the code and in the list of metrics we collect. We also have now an Ansible playbook in addition to puppet manifest. The only noticeable bug fix is about putting metrics back in the queue when we receive OSΕrror exception at the time we send graphite metrics. This is a rare event and I haven't seen it happening in production.

Changelog:

• RELEASE 0.3.15 version
• Add dcon and dses in the list of FRONTEND_METRICS
• Add ctime in the list of SERVER_AVG_METRICS
• Add chkdown in the list of server metrics
• Add slim to server metrics, in addition to frontends and backends
• Mention in README about ansible and move code under contrib
• fix update readme file
• Add Ansible Playbook For Fast Deploy haproxystats On CentOS 7.x - Update README File
• Fix a typo in README.rst
• Place metric back in the queue in case of OSError exc
• Log a message when graphite dispatcher queue is full
• Make docstrings compatible with EP257