summaryrefslogtreecommitdiffstats
path: root/HISTORICAL_CHANGELOG.md
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--HISTORICAL_CHANGELOG.md655
1 files changed, 655 insertions, 0 deletions
diff --git a/HISTORICAL_CHANGELOG.md b/HISTORICAL_CHANGELOG.md
new file mode 100644
index 000000000..3e7688f33
--- /dev/null
+++ b/HISTORICAL_CHANGELOG.md
@@ -0,0 +1,655 @@
+netdata (1.10.0) - 2018-03-27
+
+ Please check full changelog at github.
+ https://github.com/netdata/netdata/releases
+
+
+netdata (1.9.0) - 2017-12-17
+
+ Please check full changelog at github.
+ https://github.com/netdata/netdata/releases
+
+
+netdata (1.8.0) - 2017-09-17
+
+ This is mainly a bugfix release.
+ Please check full changelog at github.
+
+
+netdata (1.7.0) - 2017-07-16
+
+ * netdata is still spreading fast
+
+ we are at 320.000 users and 132.000 servers
+
+ Almost 100k new users, 52k new installations and 800k docker pulls
+ since the previous release, 4 and a half months ago.
+
+ netdata user base grows at about 1000 new users and 600 new servers
+ per day. Thank you. You are awesome.
+
+ * The next release (v1.8) will be focused on providing a global health
+ monitoring service, for all netdata users, for free.
+
+ * netdata is now a (very fast) fully featured statsd server and the
+ only one with automatic visualization: push a statsd metric and hit
+ F5 on the netdata dashboard: your metric visualized. It also supports
+ synthetic charts, defined by you, so that you can correlate and
+ visualize your application the way you like it.
+
+ * netdata got new installation options
+ It is now easier than ever to install netdata - we also distribute a
+ statically linked netdata x86_64 binary, including key dependencies
+ (like bash, curl, etc) that can run everywhere a Linux kernel runs
+ (CoreOS, CirrOS, etc).
+
+ * metrics streaming and replication has been improved significantly.
+ All known issues have been solved and key enhancements have been added.
+ Headless collectors and proxies can now send metrics to backends when
+ data source = as collected.
+
+ * backends have got quite a few enhancements, including host tags and
+ metrics filtering at the netdata side;
+ prometheus support has been re-written to utilize more prometheus
+ features and provide more flexibility and integration options.
+
+ * netdata now monitors ZFS (on Linux and FreeBSD), ElasticSearch,
+ RabbitMQ, Go applications (via expvar), ipfw (on FreeBSD 11), samba,
+ squid logs (with web_log plugin).
+
+ * netdata dashboard loading times have been improved significantly
+ (hit F5 a few times on a netdata dashboard - it is now amazingly fast),
+ to support dashboards with thousands of charts.
+
+ * netdata alarms now support custom hooks, so you can run whatever you
+ like in parallel with netdata alarms.
+
+ * As usual, this release brings dozens of more improvements, enhancements
+ and compatibility fixes.
+
+
+netdata (1.6.0) - 2017-03-20
+
+ * birthday release: 1 year netdata
+
+ netdata was first published on March 30th, 2016.
+ It has been a crazy year since then:
+
+ 225.000 unique netdata users
+ currently, at 1.000 new unique users per day
+
+ 80.000 unique netdata installations
+ currently, at 500 new installation per day
+
+ 610.000 docker pulls on docker hub
+
+ 4.000.000 netdata sessions served
+ currently, at 15.000 sessions served per day
+
+ 20.000 github stars
+
+ Thank you!
+ You are awesome!
+
+ * central netdata is here
+
+ This is the first release that supports real-time streaming of
+ metrics between netdata servers.
+
+ netdata can now be:
+
+ - autonomous host monitoring
+ (like it always has been)
+
+ - headless data collector
+ (collect and stream metrics in real-time to another netdata)
+
+ - headless proxy
+ (collect metrics from multiple netdata and stream them to another netdata)
+
+ - store and forward proxy
+ (like headless proxy, but with a local database)
+
+ - central database
+ (metrics from multiple hosts are aggregated)
+
+ metrics databases can be configured on all nodes and each node maintaining
+ a database may have a different retention policy and possibly run
+ (even different) alarms on them.
+
+ * monitoring ephemeral nodes
+
+ netdata now supports monitoring autoscaled ephemeral nodes,
+ that are started and stopped on demand (their IP is not known).
+
+ When the ephemeral nodes start streaming metrics to the central
+ netdata, the central netdata will show register them at "my-netdata"
+ menu on the dashboard.
+
+ For more information check:
+ https://github.com/netdata/netdata/tree/master/streaming#monitoring-ephemeral-nodes
+
+ * monitoring ephemeral containers and VM guests
+
+ netdata now cleans up container, guest VM, network interfaces and mounted
+ disk metrics, disabling automatically their alarms too.
+
+ For more information check:
+ https://github.com/netdata/netdata/tree/master/collectors/cgroups.plugin#monitoring-ephemeral-containers
+
+ * apps.plugin ported for FreeBSD
+
+ @vlvkobal has ported "apps.plugin" to FreeBSD. netdata can now provide
+ "Applications", "Users" and "User Groups" on FreeBSD.
+
+ * web_log plugin
+
+ @l2isbad has done a wonderful job creating a unified web log parsing plugin
+ for all kinds of web server logs. With it, netdata provides real-time
+ performance information and health monitoring alarms for web applications
+ and web sites!
+
+ For more information check:
+ https://github.com/netdata/netdata/blob/master/collectors/python.d.plugin/web_log#web_log
+
+ * backends
+
+ netdata can now archive metrics to `JSON` backends
+ (both push, by @lfdominguez, and pull modes).
+
+ * IPMI monitoring
+
+ netdata now has an IPMI plugin (based on freeipmi)
+ for monitoring server hardware.
+
+ The plugin creates (up to) 8 charts:
+
+ 1. number of sensors by state
+ 2. number of events in SEL
+ 3. Temperatures CELCIUS
+ 4. Temperatures FAHRENHEIT
+ 5. Voltages
+ 6. Currents
+ 7. Power
+ 8. Fans
+
+ It also supports alarms (including the number of sensors in critical state).
+
+ For more information, check:
+ https://github.com/netdata/netdata/tree/master/collectors/freeipmi.plugin
+
+ * new plugins
+
+ @l2isbad builds python data collection plugins for netdata at an wonderfull
+ rate! He rocks!
+
+ - **web_log** for monitoring in real-time all kinds of web server log files @l2isbad
+ - **freeipmi** for monitoring IPMI (server hardware)
+ - **nsd** (the [name server daemon](https://www.nlnetlabs.nl/projects/nsd/)) @383c57
+ - **mongodb** @l2isbad
+ - **smartd_log** (monitoring disk S.M.A.R.T. values) @l2isbad
+
+ * improved plugins
+
+ - **nfacct** reworked and now collects connection tracker information using netlink.
+ - **ElasticSearch** re-worked @l2isbad
+ - **mysql** re-worked to allow faster development of custom mysql based plugins (MySQLService) @l2isbad
+ - **SNMP**
+ - **tomcat** @NMcCloud
+ - **ap** (monitoring hostapd access points)
+ - **php_fpm** @l2isbad
+ - **postgres** @l2isbad
+ - **isc_dhcpd** @l2isbad
+ - **bind_rndc** @l2isbad
+ - **numa**
+ - **apps.plugin** improvements and freebsd support @vlvkobal
+ - **fail2ban** @l2isbad
+ - **freeradius** @l2isbad
+ - **nut** (monitoring UPSes)
+ - **tc** (Linux QoS) now works on qdiscs instead of classes for the same result (a lot faster) @t-h-e
+ - **varnish** @l2isbad
+
+ * new and improved alarms
+ - **web_log**, many alarms to detect common web site/API issues
+ - **fping**, alarms to detect packet loss, disconnects and unusually high latency
+ - **cpu**, cpu utilization alarm now ignores `nice`
+
+ * new and improved alarm notification methods
+ - **HipChat** to allow hosted HipChat @frei-style
+ - **discordapp** @lowfive
+
+ * dashboard improvements
+ - dashboard now works on HiDPi screens
+ - dashboard now shows version of netdata
+ - dashboard now resets charts properly
+ - dashboard updated to use latest gauge.js release
+
+ * other improvements
+ - thanks to @rlefevre netdata now uses a lot of different high resolution system clocks.
+
+ netdata has received a lot more improvements from many more contributors!
+
+ Thank you all!
+
+
+netdata (1.5.0) - 2017-01-22
+
+ * yet another release that makes netdata the fastest
+ netdata ever!
+
+ * netdata runs on FreeBSD, FreeNAS and MacOS !
+
+ Vladimir Kobal (@vlvkobal) has done a magnificent work
+ porting netdata to FreeBSD and MacOS.
+
+ Everyhing works: cpu, memory, disks performance, disks space,
+ network interfaces, interrupts, IPv4 metrics, IPv6 metrics
+ processes, context switches, softnet, IPC queues,
+ IPC semaphores, IPC shared memory, uptime, etc. Wow!
+
+ * netdata supports data archiving to backend databases:
+
+ - Graphite
+ - OpenTSDB
+ - Prometheus
+
+ and of course all the compatible ones
+ (KairosDB, InfluxDB, Blueflood, etc)
+
+ * new plugins:
+
+ Ilya Mashchenko (@l2isbad) has created most of the python
+ data collection plugins in this release !
+
+ - systemd Services (using cgroups!)
+ - FPing (yes, network latency in netdata!)
+ - postgres databases @facetoe, @moumoul
+ - Vanish disk cache (v3 and v4) @l2isbad
+ - ElasticSearch @l2isbad
+ - HAproxy @l2isbad
+ - FreeRadius @l2isbad, @lgz
+ - mdstat (RAID) @l2isbad
+ - ISC bind (via rndc) @l2isbad
+ - ISC dhcpd @l2isbad, @lgz
+ - Fail2Ban @l2isbad
+ - OpenVPN status log @l2isbad, @lgz
+ - NUMA memory @tycho
+ - CPU Idle @tycho
+ - gunicorn log @deltaskelta
+ - ECC memory hardware errors
+ - IPC semaphores
+ - uptime plugin (with a nice badge too)
+
+ * improved plugins:
+
+ - netfilter conntrack
+ - mysql (replication) @l2isbad
+ - ipfs @pjz
+ - cpufreq @tycho
+ - hddtemp @l2isbad
+ - sensors @l2isbad
+ - nginx @leolovenet
+ - nginx_log @paulfantom
+ - phpfpm @leolovenet
+ - redis @leolovenet
+ - dovecot @justohall
+ - cgroups
+ - disk space
+ - apps.plugin
+ - /proc/interrupts @rlefevre
+ - /proc/softirqs @rlefevre
+ - /proc/vmstat (system memory charts)
+ - /proc/net/snmp6 (IPv6 charts)
+ - /proc/self/meminfo (system memory charts)
+ - /proc/net/dev (network interfaces)
+ - tc (linux QoS)
+
+ * new/improved alarms:
+
+ - MySQL / MariaDB alarms (incl. replication)
+ - IPFS alarms
+ - HAproxy alarms
+ - UDP buffer alarms
+ - TCP AttemptFails
+ - ECC memory alarms
+ - netfilter connections alarms
+ - SNMP
+
+ * new alarm notifications:
+
+ - messagebird.com @tech-no-logical
+ - pagerduty.com @jimcooley
+ - pushbullet.com @tperalta82
+ - twilio.com @shadycuz
+ - HipChat
+ - kafka
+
+ * shell integration
+
+ - shell scripts can now query netdata easily!
+
+ * dashboard improvements:
+ - dashboard is now faster on firefox, safari, opera, edge
+ (edge is still the slowest)
+ - dashboard now has a little bigger fonts
+ - SHIFT + mouse wheel to zoom charts, works on all browsers
+ - perfect-scrollbar on the dashboard
+ - dashboard 4K resolution fixes
+ - dashboard compatibility fixes for embedding charts in
+ third party web sites
+ - charts on custom dashboards can have common min/max
+ even if they come from different netdata servers
+ - alarm log is now saved and loaded back so that
+ the alarm history is available at the dashboard
+
+ * other improvements:
+ - python.d.plugin has received way to many improvements
+ from many contributors!
+ - charts.d.plugin can now be forked to support
+ multiple independent instances
+ - registry has been re-factored to lower its memory
+ requirements (required for the public registry)
+ - simple patterns in cgroups, disks and alarms
+ - netdata-installer.sh can now correctly install
+ netdata in containers
+ - supplied logrotate script compatibility fixes
+ - spec cleanup @breed808
+ - clocks and timers reworked @rlefevre
+
+ netdata has received a lot more improvements from many more
+ contributors!
+
+ Thank you all guys!
+
+
+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/netdata/netdata/tree/master/installer#installation
+
+ In detail:
+
+ * netdata has alarms!
+
+ Based on the POLL we made on github
+ (https://github.com/netdata/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/netdata/netdata/tree/master/health
+
+ * netdata has badges!
+
+ netdata can generate badges with live information from the
+ collected metrics.
+ Please check the wiki page:
+
+ https://github.com/netdata/netdata/tree/master/web/api/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