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, [lease 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