A library to work with HAProxy via the stats socket
A Python library to manage HAProxy via stats socket.
haproxyadmin is a Python library for interacting with HAProxy load balancer to perform operations such as enabling/disabling servers. It does that by issuing the appropriate commands over the stats socket provided by HAProxy. It also uses that stats socket for retrieving statistics and changing settings.
HAProxy is a multi-process daemon and each process can only be accessed by a distinct stats socket. There isn’t any shared memory for all these processes. That means that if a frontend or backend is managed by more than one processes, you have to find which stats socket you need to send the query/command. This makes the life of a sysadmin a bit difficult as he has to keep track of which stats socket to use for a given object(frontend/backend/server).
haproxyadmin resolves this problem by presenting objects as single entities even when they are managed by multiple processes. It also supports aggregation for various statistics provided by HAProxy. For instance, to report the requests processed by a frontend it queries all processes which manage that frontend and return the sum.
The library works with Python 2.7 and Python 3.4, but for development and testing Python 3.4 is used. The Six Python 2 and 3 Compatibility Library is being used to provide the necessary wrapping over the differences between these 2 major versions of Python.
>>> from haproxyadmin import haproxy >>> hap = haproxy.HAProxy(socket_dir='/run/haproxy') >>> frontends = hap.frontends() >>> for frontend in frontends: ... print(frontend.name, frontend.requests, frontend.process_nb) ... frontend_proc2 0  haproxy 0 [4, 3, 2, 1] frontend_proc1 0  frontend1_proc34 0 [4, 3] frontend2_proc34 0 [4, 3] >>> >>> >>> backends = hap.backends() >>> for backend in backends: ... print(backend.name, backend.requests, backend.process_nb) ... servers = backend.servers() ... for server in servers: ... print(" ", server.name, server.requests) ... backend_proc2 100  bck_proc2_srv4_proc2 25 bck_proc2_srv3_proc2 25 bck_proc2_srv1_proc2 25 bck_proc2_srv2_proc2 25 haproxy 0 [4, 3, 2, 1] backend1_proc34 16 [4, 3] bck1_proc34_srv1 6 bck_all_srv1 5 bck1_proc34_srv2 5 backend_proc1 29  member2_proc1 14 member1_proc1 15 bck_all_srv1 0 backend2_proc34 100 [4, 3] bck2_proc34_srv2 97 bck2_proc34_srv1 2 bck_all_srv1 1 >>>
The documentation of the library is available at http://haproxyadmin.readthedocs.org
- HAProxy in multi-process mode (nbproc >1)
- UNIX stats socket, no support for querying HTTP statistics page
- Frontend operations
- Backend operations
- Server operations
- ACL operations
- MAP operations
- Aggregation on various statistics
- Change global options for HAProxy
sudo python setup.py install
Build (source) RPMs:
python setup.py clean --all; python setup.py bdist_rpm
Build a source archive for manual installation:
python setup.py sdist
Bump versions in docs/source/conf.py and haproxyadmin/__init__.py
Commit above change with:
git commit -av -m'RELEASE 0.1.3 version'
Create a signed tag, pbr will use this for the version number:
git tag -s 0.1.3 -m 'bump release'
Create the source distribution archive (the archive will be placed in the dist directory):
python setup.py sdist
pbr will update ChangeLog file and we want to squeeze them to the previous commit thus we run:
git commit -av --amend
Move current tag to the last commit:
git tag -fs 0.1.3 -m 'bump release'
git push;git push --tags
I would love to hear what other people think about haproxyadmin and provide feedback. Please post your comments, bug reports, wishes on my issues page.
This program was originally developed for Booking.com. With approval from Booking.com, the code was generalised and published as Open Source on github, for which the author would like to express his gratitude.