netdata (1.4.0) - 2016-10-04

At a glance:

- the fastest netdata ever (with a better look too)!

- improved IoT and containers support!

- alarms improved in almost every way!

- new plugins:

softnet netdev,

extended TCP metrics,

UDPLite

NFS v2, v3 client (server was there already),

NFS v4 server & client,

APCUPSd,

RetroShare

- improved plugins:

mysql,

cgroups,

hddtemp,

sensors,

phpfm,

tc (QoS)

In detail:

* improved alarms

Many new alarms have been added to detect common kernel

configuration errors and old alarms have been re-worked

to avoid notification floods.

Alarms now support notification hysteresis (both static

and dynamic), notification self-cancellation, dynamic

thresholds based on current alarm status

* improved alarm notifications

netdata now supports:

- email notifications

- slack.com notifications on slack channels

- pushover.net notifications (mobile push notifications)

- telegram.org notifications

For all the above methods, netdata supports role-based

notifications, with multiple recipients for each role

and severity filtering per recipient!

Also, netdata support HTML5 notifications, while the

dashboard is open in a browser window (no need to be

the active one).

All notifications are now clickable to get to the chart

that raised the alarm.

* improved IoT support!

netdata builds and runs with musl libc and runs on systems

based on busybox.

* improved containers support!

netdata runs on alpine linux (a low profile linux distribution

used in containers).

* Dozens of other improvements and bugfixes

netdata (1.3.0) - 2016-08-28

At a glance:

- netdata has health monitoring / alarms!

- netdata has badges that can be embeded anywhere!

- netdata plugins are now written in Python!

- new plugins: redis, memcached, nginx_log, ipfs, apache_cache

IMPORTANT:

Since netdata now uses Python plugins, new packages are

required to be installed on a system to allow it work.

For more information, please check the installation page:

https://github.com/firehol/netdata/wiki/Installation

In detail:

* netdata has alarms!

Based on the POLL we made on github

(https://github.com/firehol/netdata/issues/436),

health monitoring was the winner. So here it is!

netdata now has a poweful health monitoring system embedded.

Please check the wiki page:

https://github.com/firehol/netdata/wiki/health-monitoring

* netdata has badges!

netdata can generate badges with live information from the

collected metrics.

Please check the wiki page:

https://github.com/firehol/netdata/wiki/Generating-Badges

* netdata plugins are now written in Python!

Thanks to the great work of Paweł Krupa (@paulfantom), most BASH

plugins have been ported to Python.

The new python.d.plugin supports both python2 and python3 and

data collection from multiple sources for all modules.

The following pre-existing modules have been ported to Python:

- apache

- cpufreq

- example

- exim

- hddtemp

- mysql

- nginx

- phpfm

- postfix

- sensors

- squid

- tomcat

The following new modules have been added:

- apache_cache

- dovecot

- ipfs

- memcached

- nginx_log

- redis

* other data collectors:

- Thanks to @simonnagl netdata now reports disk space usage.

* dashboards now transfer a certain settings from server to server

when changing servers via the my-netdata menu.

The settings transferred are the dashboard theme, the online

help status and current pan and zoom timeframe of the dashboard.

* API improvements:

- reduction functions now support 'min', 'sum' and 'incremental-sum'.

- netdata now offers a multi-threaded and a single threaded

web server (single threaded is better for IoT).

* apps.plugin improvements:

- can now run with command line argument 'without-files'

to prevent it from enumating all the open files/sockets/pipes

of all running processes.

- apps.plugin now scales the collected values to match the

the total system usage.

- apps.plugin can now report guest CPU usage per process.

- repeating errors are now logged once per process.

* netdata now runs with IDLE process priority (lower than nice 19)

* netdata now instructs the kernel to kill it first when it starves

for memory.

* netdata listens for signals:

- SIGHUP to netdata instructs it to re-open its log files

(new logrotate files added too).

- SIGUSR1 to netdata saves the database

- SIGUSR2 to netdata reloads health / alarms configuration

* netdata can now bind to multiple IPs and ports.

* netdata now has new systemd service file (it starts as user

netdata and does not fork).

* Dozens of other improvements and bugfixes

netdata (1.2.0) - 2016-05-16

At a glance:

- netdata is now 30% faster

- netdata now has a registry (my-netdata dashboard menu)

- netdata now monitors Linux Containers (docker, lxc, etc)

IMPORTANT:

This version requires libuuid. The package you need is:

- uuid-dev (debian/ubuntu), or

- libuuid-devel (centos/fedora/redhat)

In detail:

* netdata is now 30% faster !

- Patches submitted by @fredericopissarra improved overall

netdata performance by 10%.

- A new improved search function in the internal indexes

made all searches faster by 50%, resulting in about

20% better performance for the core of netdata.

- More efficient threads locking in key components

contributed to the overal efficiency.

* netdata now has a CENTRAL REGISTRY !

The central registry tracks all your netdata servers

and bookmarks them for you at the 'my-netdata' menu

on all dashboards.

Every netdata can act as a registry, but there is also

a global registry provided for free for all netdata users!

* netdata now monitors CONTAINERS !

docker, lxc, or anything else. For each container it monitors

CPU, RAM, DISK I/O (network interfaces were already monitored)

* apps.plugin: now uses linux capabilities by default

without setuid to root

* netdata has now an improved signal handler

thanks to @simonnagl

* API: new improved CORS support

* SNMP: counter64 support fixed

* MYSQL: more charts, about QCache, MyISAM key cache,

InnoDB buffer pools, open files

* DISK charts now show mount point when available

* Dashboard: improved support for older web browsers

and mobile web browsers (thanks to @simonnagl)

* Multi-server dashboards now allow de-coupled refreshes for

each chart, so that if one netdata has a network latency

the other charts are not affected

* Several other minor improvements and bugfixes

netdata (1.1.0) - 2016-04-20

Dozens of commits that improve netdata in several ways:

- Data collection: added IPv6 monitoring

- Data collection: added SYNPROXY DDoS protection monitoring

- Data collection: apps.plugin: added charts for users and user groups

- Data collection: apps.plugin: grouping of processes now support patterns

- Data collection: apps.plugin: now it is faster, after the new features added

- Data collection: better auto-detection of partitions for disk monitoring

- Data collection: better fireqos intergation for QoS monitoring

- Data collection: squid monitoring now uses squidclient

- Data collection: SNMP monitoring now supports 64bit counters

- API: fixed issues in CSV output generation

- API: netdata can now be restricted to listen on a specific IP

- Core and apps.plugin: error log flood protection

- Dashboard: better error handling when the netdata server is unreachable

- Dashboard: each chart now has a toolbox

- Dashboard: on-line help support

- Dashboard: check for netdata updates button

- Dashboard: added example /tv.html dashboard

- Packaging: now compiles with musl libc (alpine linux)

- Packaging: added debian packaging

- Packaging: support non-root installations

- Packaging: the installer generates uninstall script

netdata (1.0.0) - 2016-03-22

- first public release

netdata (1.0.0-rc.1) - 2015-11-28

- initial packaging