Netdata just released netdata v1.15!
Netdata are very happy and proud to include two major improvements in this release, an aggregated node view and a new database engine! Also, this release contains 11 bug fixes, 30 improvements and many more.
Netdata have also added support for the AWS Kinesis backend and New collectors for:
– the Tengine web server,
– ScaleIO (VxFlex OS),
– ioping-like latency metrics,
– Energi Core node instances.
Netdata now have a new “text-only” chart type, support for cpu limits for v2 cgroups, docker swarm metrics and improved documentation.
You can read about Netdata Cloud and the future of netdata here: https://blog.netdata.cloud/posts/netdata-cloud-announcement/
Check the full release notes at https://github.com/netdata/netdata/releases
Project home page: https://github.com/netdata/netdata
Netdata is an autonomous performance and health monitoring system, unlike anything you have seen so far:
– it is a data collection agent
– is is a time-series database
– it is a super fast visualization engine
– it is a health monitoring watchdog, generating alarms
So netdata is an agent you install on all your systems (physical & virtual servers, containers, IoT).
Unlike other monitoring solutions:
– it collects unlimited metrics per server (usually a few thousands per server)
– it collects all metrics every second (1s granularity)
– it needs just 1% CPU utilization of a single core (yes! it is fast!)
– it needs about 25MB of RAM (you can configure this)
– it auto-detects most of the metrics to collect
– it comes with hundreds of alarms pre-configured
– it does not need any configuration
– it is its own web server
– it does not need any maintainance
Once installed, all your netdata installations become one virtual application. Your data do not leave the server they are collected, but your browser connects all your netdata so that you can troubleshoot multi-server performance problems without switching applications.
Netdata is very fast (faster than anything else, even commercial solutions). The time needed between collection and visualization is 1ms (1/1000 of a second, including network latencies and HTTP compression). Netdata is so effective, that it kills the console for performance troubleshooting. You can use it to trace most performance issues, without touching the console.
Netdata does not need a time-series database (prometheus, graphite, etc) or Grafana, although if you are familiar with these, you can instruct netdata to push all the metrics there too. Netdata can also re-sample the metrics and push them at a lower rate to avoid congesting these servers with data and it can filter the metrics (send fewer metrics). If you need statistics of past performance, you can use this feature for long term archiving of your data.
Netdata allows building hierarchies of servers. So each netdata can be:
– autonomous server monitoring (the default)
– headless collector (collects metrics and sends them to another netdata)
– headless proxy (receives metrics from slaves and re-transmits them)
– store and forward proxy (also presents dashboards and runs alarms)
Netdata collects metrics from all these sources:
System (CPU, RAM, swap, disks, interrupts, softirqs, softnet, entropy, IPC, page faults, huge pages, KSM, Numa, Slab, etc)
Disks (block devices, BCACHE, mdstat, hddtemp, smartd, device mapper, veritas volume manager, megacli, adaptec controller, etc)
Filesystems (BTRFS, Ceph, NFS, Samba, ZFS)
Networking (kernel network stack, IPv4 and IPv6 for all protocols, netfilter/iptables, SYNPROXY, NFACCT, QoS/tc, access points and hostapd, OpenVPN, LibreSwan, fping, port check)
Processes (system processes, applications, systemd services)
Users (by grouping the process tree, logind)
Containers (systemd-nspawn, lxc, lxd, docker, kubernetes)
Libvirt VMs (qemu)
Web servers (apache, nginx, nginx+, lighttpd, ipfs, litespeed, phpfpm, tomcat, web server access.log files, http check)
Proxies and Balancers (haproxy, squid, traefik, varnish, IPVS)
Databases servers (couchdb, memcached, mongodb, mysql, mariadb, postresql, redis, rethinkdb)
Message queues (rabbitmq, beanstalkd)
Search and indexing (elasticsearch)
DNS servers (bind / named, nsd, powerdns, unbound, DNS query time)
Time servers (chrony, ntpd)
Mail servers (dovecot, postfix, exim)
Hardware sensors (IPMI, lm-sensors, RPi, w1sensor)
UPSes (apcupsd, nut)
Telephony servers (opensips)
Provisioning systems (puppet)
and dozens more! You can also extend it, by writing plugins that collect data from any source, using any computer language.
Netdata can also be used for APM (Application Performance Monitoring). It is a fully featured (and very fast) statsd server, it can collect metrics from Go application using the expvar package and Java Spring Boot applications.
Netdata supports the following alarm notification methods: alerta.io, AWS SNS, discordapp.com, email, fleep.io, flock.com, IRC channels, kavenegar.com, messagebird.com, pagerduty.com, pushbullet.com, pushover.net, rocket.chat, slack.com, syslog, telegram.org, twilio.com and execute custom scripts.
Netdata has been designed by SysAdmins, DevOps and Developers to help them get immediate results, not just visualize metrics.
Netdata is free, open-source software (FOSS). It runs on Linux, FreeBSD and MacOS.
If you like it, just give Netdata to GitHub star…
Project home page: https://github.com/netdata/netdata