Netdata v1.18.1

Release v1.18.1 contains 17 bug fixes, 5 improvements, and 5 documentation updates.

At a glance

Patch release 1.18.1 contains several bug fixes, mainly related to FreeBSD and the binary package generation process.

Netdata can now send notifications to Google Hangouts Chat!

On certain systems, the slabinfo plugin introduced in v1.18.0 added thousands of new metrics. We decided the collector's usefulness to most users didn't justify the increase in resource requirements. This release disables the collector by default.

Finally, we added a chart under Netdata Monitoring to present a better view of the RAM used by the database engine (dbengine). The chart doesn't currently take into consideration the RAM used for slave nodes, so we intend to add more related charts in the future.

Acknowledgements

We'd like to thank:

• hendrikhofstadt for the Google Hangouts notifications
• stevenh for the awesome zombie process reaper and the fix for the freeipmi collector
• samm-git for the addition of the VMware VMXNET3 driver to the default interfaces list for FreeBSD
• sz4bi for a documentation fix

Improvements

• Disable slabinfo plugin by default to reduce the total number of metrics collected #7056 (vlvkobal)
• Add dbengine RAM usage statistics #7038 (mfundul)
• Support Google Hangouts chat notifications #7013 (hendrikhofstadt)
• Add CMocka unit tests #6985 (vlvkobal)
• Add prerequisites to enable automatic updates for installations via the static binary (kickstart-static64.sh) #7060 (knatsakis)

Documentation

• Fix typo in health_alarm_notify.conf #7062 (sz4bi)
• Fix BSD/pfSense documentation #7041 (thiagoftsm)
• Document the structure of the api/v1/data API responses. #7012 (amoss)
• Tutorials to support v1.18 features #6993 (joelhans)
• Fix broken links in docs #7123 (joelhans)

Bug fixes

• Fix unbound collector timings: Convert recursion timings to milliseconds. #7121 (Ferroin)
• Fix unbound collector unhandled exceptions #7112 (ilyam8)
• Fix upgrade path from v1.17.1 to v1.18.x for deb packages #7118 (knatsakis)
• Fix CPU charts in apps plugin on FreeBSD #7115 (vlvkobal)
• Fix megacli collector binary search and sudo check #7108 (ilyam8)
• Fix missing packages, by running the triggers for DEB and RPM package build in separate stages #7105 (knatsakis)
• Fix segmentation fault in FreeBSD when statsd is disabled #7102 (vlvkobal)
• Fix Clang warnings #7090 (thiagoftsm)
• Fix python.d error logging: change chart suppress msg level from ERROR to INFO #7085 (ilyam8)
• Fix freeipmi update frequency check: was warning that 5 was too frequent and it was setting it to 5. #7078 (stevenh)
• Fix alarm configurations not getting loaded, via better handling of chart names with special characters #7069 (thiagoftsm)
• Fix dbengine not working when mmap fails - mostly with BSD kernels #7065 (mfundul)
• Fix FreeBSD issue due to incorrect size of a zeroed block #7061 (vlvkobal)
• Don't write HTTP response 204 messages to the logs #7035 (vlvkobal)
• Fix build when CMocka isn't installed #7129 (vlvkobal)
• FreeBSD plugin: Add VMware VMXNET3 driver to the default interfaces list #7109 (samm-git)
• Prevent zombie processes when a child is re-parented to netdata when its running in a container , by adding child process reaper #7059 (stevenh)

Netdata v1.18.0

Release v1.18.0 contains 5 new collectors, 19 bug fixes, 28 improvements, and 20 documentation updates.

At a glance

The database engine is now the default method of storing metrics in Netdata. You immediately get more efficient and configurable long-term metrics storage without any work on your part. By saving recent metrics in RAM and "spilling" historical metrics to disk for long-term storage, the database engine is laying the foundation for many more improvements to distributed metrics.

We even have a tutorial on switching to the database engine and getting the most from it. Or, just read up on how performant the database engine really is.

Both our python.d and go.d plugins now have more intelligent auto-detection by periodically dump a list of active modules to disk. When Netdata starts, such as after a reboot, the plugins use this list of known services to re-establish metrics collection much more reliably. No more worrying if the service or application you need to monitor starts up minutes after Netdata.

Two of our new collectors will help those with Hadoop big data infrastructures. The HDFS and Zookeeper collection modules come with essential alarms requested by our community and Netdata's auto-detection capabilities to keep the required configuration to an absolute minimum. Read up on the process via our HDFS and Zookeeper tutorial.

Speaking of new collectors—we also added the ability to collect metrics from SLAB cache, Gearman, and vCenter Server Appliances.

Before v1.18, if you wanted to create alarms for each dimension in a single chart, you need to write separate entities for each dimension—not very efficient or user-friendly. New dimension templates fix that hassle. Now, a single entity can automatically generate alarms for any number of dimensions in a chart, even those you weren't aware of! Our tutorial on dimension templates has all the details.

v1.18 brings support for installing Netdata on offline or air-gapped systems. To help users comply with strict security policies, our installation scripts can now install Netdata using previously-downloaded tarball and checksums instead of downloading them at runtime. We have guides for installing offline via kickstart.sh or kickstart-static64.sh in our installation documentation. We're excited to bring real-time monitoring to once-inaccessible systems!

Acknowledgements

Our thanks go to:

• Saruspete for several contributions, including the new slabinfo collector, that monitors SLAB cache mechanism metrics.
• agronick for the new Gearman worker statistics
  collector
• OneCodeMonkey for a bug fix in the alarm notification script.
• lets00 for providing a Portuguese (Brazil) translation of the installation instructions
• mbarper and davent for improvements to the uninstaller.
• n0coast for a documentation fix.

Improvements

Database engine

• Make dbengine the default memory mode #6977 (mfundul)
• Increase dbengine default cache size #6997 (mfundul)
• Reduce overhead during write IO #6964 (mfundul)
• Detect deadlock in dbengine page cache #6911 (mfundul)
• Remove hard cap from page cache size to eliminate deadlocks. #7006 (mfundul)

New Collectors

• SLAB cache mechanism (Saruspete)
• Gearman worker statistics
• vCenter Server Appliance
• Zookeeper servers
• [Hadoop Distributed File System (HDFS) nodes] (https://docs.netdata.cloud/collectors/go.d.plugin/modules/hdfs/)

Collector improvements

• rabbitmq: Add vhosts message metrics from /api/vhosts #6976 (ilyam8)
• elasticsearch: collect metrics from _cat/indices #6965 (ilyam8)
• mysql: collect galera cluster metrics #6962 (ilyam8)
• Allow configuration of the python.d launch command from netdata.conf #6781 (amoss)
• x509check: smtp cert check support (netdata/go.d.plugin#261)
• dnsmasq_dhcp: respect conf-dir,conf-file,dhcp-host options (netdata/go.d.plugin#268)
• plugin: respect previously running jobs after plugin restart (#6499)
• httpcheck: add current state duration chart (netdata/go.d.plugin#270 )
• springboot2: fix context (netdata/go.d.plugin#263)

Health

• Enable alarm templates for chart dimensions #6560 (thiagoftsm)
• Center the chart on the proper chart and time whenever an alarm link is clicked #6391 (thiagoftsm)

Installation/Packages

• netdata/installer: Add support for offline installations using kickstart.sh or kickstart-static64.sh #6693 (paulkatsoulakis)
• Allow netdata service installation, when docker runs systemd #6987 (paulkatsoulakis)
• Make spec file more consistent with version dependencies #6948 (paulkatsoulakis)
• Fix broken links on web files, for DEB #6930 (paulkatsoulakis)
• Introduce separate CUPS package for DEB #6724 and RPM #6700 distributions. (paulkatsoulakis). Do not build CUPS plugin subpackage on CentOS 6 and CentOS 7 #6926 (knatsakis)
• Various Improvements in the package release CI/CD flow #6914 #6905 #6842 #6837 #6838 #6834 (paulkatsoulakis), #6900 (cakrit)
• Remove RHEL7 - i386 binary distribution, until bug #6849 is resolved #6902 (paulkatsoulakis)
• Bring on board two scripts that build libuv and judy from source #6850 (paulkatsoulakis)

Documentation

• Add Portuguese (Brazil) translation of the installation instructions #16(lets00), #7004 (cakrit)
• Fix broken links found via linkchecker #6983 (joelhans)
• Clarification on configuring notification recipients #6961 (cakrit)
• Fix Remark Lint for READMEs in database #6942, contrib #6921, daemon README #6920 and backends #6917 (prhomhyse)
• Suggest using /run or /var/run for the unix socket #6916 (cakrit)
• Improve documentation for the SNMP collector #6915 (cakrit)
• Update docs for offline install #6884 (paulkatsoulakis)
• Remove Dollar sign from Bash code in documentation and fix remark-lint warnings #6880 (prhomhyse)
• Markdown syntax fixes for MDX parser #6877 (joelhans)
• Update python.d module checklist to match the current paths and build system. #6874 (Ferroin)
• Add instructions for simple SMTP transport #6870 (cakrit)
• Add example for prometheus archiving source parameter #6869 (cakrit)
• Fix broken links in the standard web dashboard doc #6854 (prhomhyse)
• Overhaul of Getting started guide #6811 (joelhans)
• NPM Packages version update #6801 (prhomhyse)
• Update suggested grep command in “high performance netdata” to be more specific #6794 (n0coast)

Other

• API: Include family into the allmetrics JSON response #6966 (ilyam8)
• API: Add fixed width option to badges #6903 (underhood)
• Allow hostnames in Access Control Lists #6796 (amoss)
• Functional test improvements for web and alarms tests #6783 (thiagoftsm)

Bug fixes

• Fix issue error in alarm notification script, when executed without any arguments #7003 (OneCodeMonkey)
• Fix Coverity warnings #6992 #6970 #6941 #6797 (thiagoftsm), #6909 (cakrit)
• Fix dbengine consistency when a writer modifies a page concurrently with a reader querying its metrics #6979 (mfundul)
• Fix memory leak on netdata exit #6945 (vlvkobal)
• Fix for missing boundary data points in certain cases #6938 (mfundul)
• Fix unhandled exception log warnings in the python.d collector orchestrator start\_job #6928 (ilyam8)
• Fix CORS errors when accessing the health management API, by permitingt x-auth-token in Access-Control-Allow-Headers #6894 (cakrit)
• Fix misleading error log entries RRDSET: chart name 'XXX' on host 'YYY' already exists, by changing the log level for chart updates #6887 (vlvkobal)
• Properly resolve all Kubernetes container names #6885 (cakrit)
• Fix LGTM warnings #6875 (jacekkolasa)
• Fix agent UI redirect loop during cloud sign-in #6868 (jacekkolasa)
• Fix /var/lib/netdata/registry getting left behind after uninstall #6867 (davent)
• Fix python.d.plugin bug in parsing configuration files with no explicitly defined jobs #6856 (ilyam8)
• Fix potential buffer overflow in the web server #6817 (amoss)
• Fix netdata group deletion on linux for uninstall script #6645 (mbarper)
• Various cppcheck fixes #6386 (ac000)
• Fix crash on FreeBSD due to do_dev_cpu_temperature stack corruption #7014 (samm-git)
• Fix handling of illegal metric timestamps in database engine #7008 (mfundul)
• Fix a resource leak #7007 (vlvkobal)
• Fix rabbitmq collector error when no vhosts are available. #7018 (mfundul)

Netdata v1.17.1

Release v1.17.1 contains 2 bug fixes, 6 improvements, and 2 documentation updates.

At a glance

The main reason for the patch release is an essential fix to the repeating alarm notifications we introduced in v1.17.0. If you enabled repeating notifications, Netdata would not then send CLEAR notifications for the selected alarms.

The release also includes a significant improvement to Netdata's auto-detection capabilities, especially after a system restart. Netdata now remembers which python.d plugin jobs were successfully collecting data the last time it was running, and retries to run those jobs for 5 minutes before giving up. As a result, you no longer have to worry if your system starts Netdata before the monitored services have had a chance to start properly. We will complete the same improvement for go.d plugins in v1.18.0.

We also made some improvements to our binary packages and added a neat sample custom dashboard that can show charts from multiple Netdata agents.

Acknowledgements

Our thanks go to:

• tnyeanderson for Dash.html, the custom dashboard that can show charts from multiple hosts.
• qingkunl for improving the charts auto-scaling feature with nanosec and num units.
• Fohdeesha for documentation improvements
• Saruspete for improving debugging capabilities with tags for threads and his significant involvement in many other issues

Improvements

Binary packages

• netdata/packaging: Trigger stable package generation upon release process #6766 (paulkatsoulakis)
• netdata/packaging: Fix ubuntu/xenial runtime dependencies #6825 (paulkatsoulakis)
• netdata/packaging: Remove fedora/28, which is no longer available #6808 (paulkatsoulakis)
• netdata/packaging: Override control file for debian/buster #6777 (paulkatsoulakis)

GUI

• Expand dashboard auto-scaling and convertible units. Added two more units that allow auto-scaling and conversion: nanoseconds and num. #5920 (qingkunl)

Collector improvements

• Auto-detect previously running python.d jobs and retry for 5 minutes #6661 (ilyam8)

Documentation

• Fix pfsense instructions and links #6768 (Fohdeesha)
• Add high level explanation of dashboard contents #6648 (joelhans)

Other

• Update cache hashes for js and css #6756 (jacekkolasa)
• Provide a tag to identify the thread in the error messages. #6745 (Saruspete)
• Add sample multi-server dashboard dash.html #6603 (tnyeanderson)
• Replace hard-coded HTTP response codes #6595 (thiagoftsm)

Bug fixes

• Fix clear notifications for repeating alarms #6638 (thiagoftsm)
• Stop configure.ac from linking against dbengine and https libraries when dbengine or https are disabled #6658 (mfundul)

Release v1.17.0 contains 38 bug fixes, 33 improvements, and 20 documentation updates.

At a glance

You can now change the data collection frequency at will, without losing previously collected values. A major improvement to the new database engine allows you not only to store metrics at variable granularity, but also to autoscale the time axis of the charts, depending on the data collection frequencies used during the presented time.

You can also now monitor VM performance from one or more vCenter servers with a new VSphere collector. In addition, the proc plugin now also collects ZRAM device performance metrics and the apps plugin monitors process uptime for the defined process groups.

Continuing our efforts to integrate with as many existing solutions as possible, you can now directly archive metrics from Netdata to MongoDB via a new backend.

Netdata badges now support international (UTF8) characters! We also made our URL parser smarter, not only for international character support, but also for other strange API queries.

We also added .DEB packages to our binary distribution repositories at Packagecloud, a new collector for Linux zram device metrics, and support for plain text email notifications.

This release includes several fixes and improvements to the TLS encryption feature we introduced in v1.16.0. First, encryption slave-to-master streaming connections wasn't working as intended. And second, our community helped us discover cases where HTTP requests were not correctly redirected to HTTPS with TLS enabled. This release mitigates those issues and improves TLS support overall.

Finally, we improved the way Netdata displays charts with no metrics. By default, Netdata displays charts for disks, memory, and networks only when the associated metrics are not zero. Users could enable these charts permanently using the corresponding configuration options, but they would need to change more than 200 options. With this new improvement, users can enable all charts with zero values using a single, global configuration parameter.

Acknowledgements

Our thanks go to:

• Steve8291 for all his help across the board!
• alpes214 for improvements in health monitoring
• fun04wr0ng for fixing a bug in the nfacct plugin
• RaZeR-RBI for the ZRAM collector module
• underhood for the UTF-8 parsing fixes in badges, that gave us support for internationalized badges
• Ferroin](https://github.com/Ferroin) for improving the python.d collectors handling of disconnected sockets
• dex4er for improving our OS detection code
• knatsakis for his help in our CI/CD pipeline
• sunflowerbofh for .gitignore fixes
• Cat7373 for fixing some issues with the spigotmc collector

Improvements

Database engine

• Variable granularity support for data collection #6430 (mfundul)
• Added tips on the UI to encourage users to try the new DB Engine, when they reach the end of their metrics history #6711 (jacekkolasa)

Binary packages

• Added nightly generation of RPM/DEB amd64 packages #6675 (paulkatsoulakis)
• Provided built-in support for the prometheus remote write API in our packages #6480 (paulkatsoulakis)
• Documented distribution support matrix and functionality availability #6552 (paulkatsoulakis)

Health

• Added support for plain text only email notifications #6485 (leo-lb)
• Started showing “hidden” alarm variables in the responses of the chart and data API calls (#6054) #6615 (alpes214)
• Added a new API call for alarm status counters, as a first step towards badges that will show the total number of alarms #6554 (alpes214)

Security

• Added configurable default locations for trusted CA certificates #6549 (thiagoftsm)
• Added safer way to get container names #6441 (ViViDboarder)
• Added SSL connection support to the python mongodb collector #6546 (ilyam8)

New collectors

• VSphere collector go.d.plugin PR241 #6572 (ilyam8)

Collector improvements

• rethinkdb collector new driver support #6431 (ilyam8)
• The apps plugin now displays process uptime charts #6654 (vlvkobal)
• Added ZRAM device metrics to the proc.plugin #6276 #6424 (RaZeR-RBI)

Archiving

• Added a new MongoDB backend #6524 (vlvkobal)

Documentation

• Add a statement about permissions for the diskspace plugin #6474 (vlvkobal)
• Improved the running behind Nginx guide #6466 (prhomhyse)
• Add more supported backends to the documentation #6443 (vlvkobal)
• Removed Ventureer from the list of demo sites #6442 (paulkatsoulakis)
• Updated docs health monitoring and health management api documentation #6435 (jghaanstra)
• Fixed issues in HTML docs generation, causing the hyperlink checks to function improperly #6433 (cakrit)
• New 'homepage' for documentation site #6428 (joelhans)
• Styling improvements to documentation #6425 (joelhans)
• Add documentation for binary packages, plus draft table for distributions support #6422 (paulkatsoulakis)
• Update netdata installation dependencies #6421 (paulkatsoulakis)
• Added better explanation of nightly and stable releases #6388 (joelhans)
• Add netdata haproxy documentation page #6454 (johnramsden)
• Added Netdata Cloud documentation #6476 (joelhans)
• Removed text about nightly version #6534 (joelhans)
• Provided documentation style guide & build instructions #6563 (joelhans)
• Install Netdata with Docker #6596 (prhomhyse)
• Fixed typos in: 'README.md' file. #6604 (coffeina)
• Change "netdata" to "Netdata" in all docs #6621 (joelhans)
• Fixed Markdown Lint warnings #6664 (prhomhyse)
• Improved Apache reverse proxy documentation on Content Security Policy #6667 (sunflowerbofh)

Other

• Updated our CLA, clarifying our intention to keep netdata FOSS #6504 (cakrit)
• Updated terms of use for U.S. legal reasons #6631 (cakrit)
• Updated logos in the infographic and remaining favicons #6417 (cakrit)
• SSL vs. TLS consistency and clarification in documentation #6414 (joelhans)
• Update Running-behind-apache.md #6406 (Steve8291)
• Fix Web API Health documentation #6404 (thiagoftsm)
• Added apps grouping debug messages #6375 (vlvkobal)
• GCC warning and linting improvements #6392 (ac000)
• Minor code readability changes #6539 (underhood)
• Added global configuration option to show charts with zero metrics #6419 (vlvkobal)
• Improved the way we parse HTTP requests, so we can avoid issues from edge cases #6247 #6714 (thiagoftsm)
• Build DEB and RPM packages in parallel #6579 (knatsakis)
• Updated package version requirements for LZ4 and libuv #6607 (mfundul)
• Improved system OS detection for RHEL6 and Mac OS X #6612 (dex4er)
• .travis.yml: Remove 'sudo: true' as it is now deprecated #6624 (knatsakis)
• Modified the documentation build process to accept <> around links in markdown #6646 (cakrit)
• Fixed spigotmc module typos in comments. #6680 (Cat7373)

Bug fixes

• Fixed the snappy library detection in some versions of OpenSuSE and CentOS #6479 (vlvkobal)
• Fixed sensor chips filtering in python sensors collector #6463 (ilyam8)
• Fixed user and group names in apps.plugin when running in a container, by mounting and reading /etc/passwd #6472 (vlvkobal)
• Fixed possible buffer overflow in the JSON parser used for health notification silencers #6460 (thiagoftsm)
• Fixed handling of corrupted DB files in dbengine, that could cause netdata to not start properly (CRC and I/O error handling) #6452 (mfundul)
• Stopped docs icon from linking to streaming page instead of docs root #6445 (joelhans)
• Fixed an issue with Netdata snapshots that could sometimes cause a problem during import. #6400 (jacekkolasa)
• Fixed bug that would cause netdata to attempt to kill already terminated threads again, on shutdown. #6387 (emmrk)
• Fixed out of memory (12) errors by reimplementing the myopen() function family #6339 (mfundul)
• Fixed wrong redirection of users signing in after clicking Nodes #6544 (jacekkolasa)
• Fixed python.d smartd collector increasing CPU usage #6540 (ilyam8)
• Fixed missing navigation arrow in Documentation #6533 (joelhans)
• Fixed mongodb python collector stock configuration mistake, by changing password to pass #6518 (ilyam8)
• Fixed broken left navbar links in translated docs #6505 (cakrit)
• Fixed handling of UTF8 characters in badges and added International Support to the URL parser #6426 (underhood)
• Fixed nodes menu sizing (responsive) #6455 (builat)
• Fixed issues with http redirection to https and streaming encryption #6468 (thiagoftsm)
• Fixed broken links to arcstat.py and arc_summary.py in dashboard_info.js #6461 (TheLovinator1)
• Fixed bug with the nfacct plugin that resulted in missing dimensions from the charts #6098 (fun04wr0ng)
• Stopped anonymous stats from trying to write a log under /tmp #6491 (cakrit)
• Fixed a problem with edit-config, the configuration editor, not being able to run in MacOS. We no longer deliver edit-config as part of the distribution tarball, so that it can get generated with proper configuration during installation .#6507 (paulkatsoulakis)
• Fixed issue with the netdata-updater that caused it not to run properly in static64 installations. #6520 (paulkatsoulakis)
• Fixed some yamllint errors in our Travis configuration #6526 (knatsakis)
• Properly delete obsolete dimensions for inactive disks in smartd_log #6547 (ilyam8)
• Fixed .environment file getting overwritten, by moving tarball checksum information into lib dir of netdata #6555 (paulkatsoulakis)
• Fixed handling of disconnected sockets in unbound python.d collector. #6561 (Ferroin)
• Fixed crash in malloc #6583 (thiagoftsm)
• Fixed installer error undefined reference to LZ4_compress_default #6589 (mfundul)
• Fixed issue with mysql collector that resulted in showing only a single slave_status chart, regardless of the number of replication channels #6597 (ilyam8)
• Fixed installer issue that would automatically enable the netdata service, even, if it was previously disabled #6606 (paulkatsoulakis)
• Fixed a segmentation fault in backends #6627 (vlvkobal)
• Fixed spigotmc plugin bugs #6635 (Cat7373)
• Fixed installer error when running kickstart.sh as a non-privileged user #6642 (paulkatsoulakis)
• Fixed issue causing OpenSSL libraries to not be found on gentoo #6670 (paulkatsoulakis)
• Fixed dbengine 100% CPU usage due to corrupted transaction payload handling #6731 (mfundul)
• Fixed wrong default paths in certain installations #6678 (paulkatsoulakis)
• Fixed exact path to netdata.conf in .gitignore #6709 (sunflowerbofh)
• Fixed static64 installer bug that resulted in always overwriting configuration #6710 (paulkatsoulakis)

Thanks to the community for their help!

Release v1.16.0 contains 40 bug fixes, 31 improvements and 20 documentation updates

At a glance

Binary distributions. To improve the security, speed and reliability of new netdata installations, we are delivering our own, industry standard installation method, with binary package distributions. The RPM binaries for the most common OSs are already available on packagecloud and we’ll have the DEB ones available very soon. All distributions are considered in Beta and, as always, we depend on our amazing community for feedback on improvements.

• Our stable distributions are at netdata/netdata @ packagecloud.io
• The nightly builds are at netdata/netdata-edge @ packagecloud.io

Netdata now supports SSL encryption! You can secure the communication to the web server, the streaming connections from slaves to the master and the connection to an openTSDB backend.

This version also brings two long-awaited features to netdata’s health monitoring:

• The health management API introduced in v1.12 allowed you to easily disable alarms and/or notifications while netdata was running. However, those changes were not persisted across netdata restarts. Since part of routine maintenance activities may involve completely restarting a monitoring node, netdata now saves these configurations to disk, every time you issue a command to change the silencer settings. The new LIST command of the API allows you to view at any time which alarms are currently disabled or silenced.
• A way for netdata to repeatedly send alarm notifications for some, or all active alarms, at a frequency of your choosing. As a result, you will no longer have to worry about missing a notification, forgetting about a raised alarm. The default is still to only send a single notification, so that existing users are not surprised by a different behavior.

As always, we’ve introduced new collectors, 5 of them this time.

• Of special interest to people with Windows servers in their infrastructure is the WMI collector, though we are fully aware that we need to continue our efforts to do a proper port to Windows.
• The new perf plugin collects system-wide CPU performance statistics from Performance Monitoring Units (PMU) using the perf_event_open() system call. You can read a wonderful article on why this is useful here.
• The other three are collectors to monitor Dnsmasq DHCP leases, Riak KV servers and Pihole instances.

Finally, the DB Engine introduced in v1.15.0 now uses much less memory and is more robust than before.

Acknowledgements

As you’ll see in the detailed list below, once again we’ve had great help from our contributors.

• Steve8291 was helping everywhere
• apardyl added useful new alarms and helped with documentation
• jchristgit wrote the Riak KV collector
• Saruspete made improvements to the freeipmi plugin
• kam1kaze has added new charts to the python mysql collector
• akwan and mbarper improved the application monitoring, with new process groupings
• nodiscc helped with bug and documentation fixes
• dankohn) helped with the documentation
• andvgal added an amazing configuration to help us run proper lint checks on our markdown files
• octomike, Danamir, mbarper, Wing924, n0coast and toofar delivered bug fixes
• josecv helped improve the Kubernetes helm chart.

We can't stress enough the immense help we get just from users creating an issue in GitHub, helping us identify the root cause and validate the change in their infrastructure. Unfortunately, we are not able to list all of them here, but their contribution is invaluable.

Improvements

Binary packages

• Introduced automatic binary packages generation and delivery for RPM types (Phase 1) #6223 #6369 (paulkatsoulakis)

Health

• Easily disable alarms, by persisting the silencers configuration #6274 #6360 (thiagoftsm)
• Repeating alarm notifications #6309 (thiagoftsm) and (kamcpp)
• Simplified the health cmdapi tester - no setup/cleanup needed #6210 (cakrit)
• Αdd last_collected alarm to the x509check collector #6139 (ilyam8)
• New alarm for abnormally high number of active processes. #6116 (apardyl)

Security

• SSL support in the web server and streaming/replication #5956 (thiagoftsm)
• Support encrypted connections to OpenTSDB backends #6220 (thiagoftsm)
• Show the security policy directly from GitHub #6163 #6166 (cakrit)

New collectors

• Go.d collector modules for WMI, [Dnsmasq DHCP leases)(https://github.com/netdata/go.d.plugin/tree/master/modules/dnsmasq_dhcp) and Pihole (ilyam8)
• Riak KV instances collector #6286 (jchristgit)
• CPU performance statistics using Performance Monitoring Units (PMU) via the perf_event_open() system call. (perf plugin) #6225 (vlvkobal)

Collector improvements

• Handle different sensor IDs for the same element in the freeipmi plugin #6296 (Saruspete)
• Increase the cpu_limit chart precision in cgroup plugin #6172 (vlvkobal)
• Added userstats and deadlocks charts to the python mysql collector #6118 #6115 (kam1kaze)
• Add perforce server process monitoring to the apps plugin #6064 (akwan)

Backends

• Prometheus remote write backend #6062 (vlvkobal)

DB engine improvements

• Reduced memory requirements by 40-50% #6134 (mfundul)
• Reduced the number of pages needed to be stored and indexed when using memory mode = dbengine, by adding empty page detection #6173 (mfundul)

Rebranding

• Updated the netdata logo and changed links to point to the new website #6359 #6398 (cakrit), #6396 (ivorjvr), #6389 (joelhans)

Documentation

• Improve documentation about file descriptors and systemd configuration. #6372 (mfundul)
• Update the documentation on charts with zero metrics #6314 (vlvkobal)
• Document that that in versions before 1.16, the plugins.d directory may be installed in a different location in certain OSs #6301 (cakrit)
• Remove single and multi-threaded web server configuration instructions #6291 (nodiscc)
• Add more info on the stream.conf option health enabled by default = auto #6281 (cakrit)
• Add comments about AWS SDK for C++ installation #6277 (vlvkobal)
• Fix on the installation readme regarding the supported systems (first came RedHat, then the others) #6271 (paulkatsoulakis)
• Update the new dbengine documentation #6264 (mfundul)
• Remove CNCF logo and TOC presentation reference #6234 (dankohn)
• Added code style guidance to CONTRIBUTING #6212 (cakrit)
• Visibility fix for anonymous statistics #6208 (cakrit)
• smartd documentation improvements #6207 (cakrit), #6203 (Steve8291)
• Made custom notification's instructions clearer #6181 (cakrit)
• Fix typo in the web server README #6146 (cakrit)
• Registry documentation fixes #6144 (cakrit)
• Changed 'netdata' to 'Netdata' in /docs/ and /README.md #6137 (apardyl)
• Update installer readme with OpenSUSE dependencies #6111 (mfundul)
• Fixed minor typos in the daemon configuration documentation #6090 (Steve8291)
• Mention anonymous statistics in additional places in the docs #6084 (cakrit)
• Local remark-lint checks and autofix support #5898 (andvgal)

Other

• Pass the the cloud base url parameter to the notifications mechanism, so that modifications to the configuration are respected when creating the link to the alarm #6383 (ladakis)
• Added a .gitattributes file to improve git diff for C files #6381 (ac000)
• Improved logging, to be able to trace the CRITICAL: main[main] SIGPIPE received. error #6373 (vlvkobal)
• Modify the limits of the stale bot, to close stale questions/discussions in GitHub faster #6297 (ilyam8)
• Internal CI/CD improvements #6282 #6268 (paulkatsoulakis)
• netdata/packaging: Add more distribution validations #6235 (paulkatsoulakis)
• Move call to send_statistics later, to get more telemetry events from docker containers #6113 (vlvkobal), #6096 (cakrit)
• Use github templating mechanisms to classify issues when they are created #5776 (paulfantom)

Bug fixes

• Fixed ram_available alarm #6261 (octomike)
• Stop monitoring /dev and /run in the disk space and inode usage charts #6399 (vlvkobal)
• Fixed the monitoring of the “time” group of processes #6397 (mbarper)
• Fixed compilation error PERF_COUNT_HW_REF_CPU_CYCLES' undeclared here in old Linux kernels (perf plugin) #6382 (vlvkobal)
• Fixed autodetection for openldap on Debian (apps.plugin) #6364 (nodiscc)
• Fixed compilation error on CentOS 6 (nfacct plugin) #6351 (vlvkobal)
• Fixed invalid XML page error (tomcat plugin) #6345 (Danamir)
• Remove obsolete monit metrics #6340 (ilyam8)
• Fixed Failed to parse error in adaptec_raid #6338 (ilyam8)
• Fixed cluster_health_nodes and cluster_stats_nodes charts in the elasticsearch collector #6311 (Wing924)
• A modified slave chart's "name" was not properly transferred to the master (streaming) #6304 (vlvkobal)
• Netdata could run out of file descriptors when using the new DB engine #6303 (mfundul)
• Fixed UI behavior when pressing the End key #6294 (thiagoftsm)
• Fixed UI link to check the configuration file, to open in a new tab #6294 (thiagoftsm)
• Fixed files not found during installation, due to different than expected location of the libexecdir directory #6272 (paulkatsoulakis)
• Prevented Error: 'module' object has no attribute 'Retry' messages from python collectors, by enforcing minimum version check for the UrlService library #6263 (ilyam8)
• Fixed typo that causes nfacct.plugin log messages to incorrectly show freeipmi #6260 (vlvkobal)
• Fixed netdata/netdata docker image failure, when users pass a PGID that already exists on the system #6259 (paulkatsoulakis)
• The daemon could get stuck during collection or during shutdown, when using the new dbengine. Reduced new dbengine IO utilization by forcing page alignment per dimension of chart. #6240 (mfundul)
• Properly handle timeouts/no response in dns_query_time python collector #6237 (n0coast)
• When a collector restarted after having stopped for a long time, the new dbengine would consume a lot of CPU resources. #6216 (mfundul)
• Fixed error Assertion old_state & PG_CACHE_DESCR_ALLOCATED' failed` of the new dbengine. Eliminated a page cache descriptor race condition #6202 (mfundul)
• tv.html failed to load the three left charts when accessed via https. Turn tv.html links to https #6198 (cakrit)
• Change print level from error to info for messages about clearing old files from the database#6195 (mfundul)
• Fixed warning regarding the x509check_last_collected_secs alarms. Changed the template update frequency to 60s, to match the chart’s update frequency #6194 (ilyam8)
• Email notification header lines were not terminated with \r\n as per the RFC #6187 (toofar)
• Some log entries would not be caught by the python web_log plugin. Fixed the regular expressions #6138 #6180 (ilyam8)
• Corrected the date used in pushbullet notifications #6179 (cakrit)
• Fixed FATAL error when using the new dbengine with no direct I/O support, by falling back to buffered I/O #6174 (mfundul)
• Fixed compatibility issues with varnish v4 (varnish collector) #6168 (ilyam8)
• The total number of disks in mdstat.XX_disks chart was displayed incorrectly. Fixed the "inuse" and "down" disks stacking. #6164 (vlvkobal)
• The config option --disable-telemetry was being checked after restarting netdata, which means that we would still send anonymous statistics the first time netdata was started. #6127 (cakrit)
• Fixed apcupsd collector errors, by passing correct info to the run function. #6126 (Steve8291)
• apcupsd and libreswan were not enabled by default #6120 (Steve8291)
• Fixed incorrect module name: energi to energid #6112 (Steve8291)
• The nodes view did not work properly when a reverse proxy was configured to access netdata via paths containing subpaths (e.g. myserver/netdata) #6093 (gmosx)
• Fix error message PLUGINSD : cannot open plugins directory #6080 #6089 (Steve8291)
• Corrected invalid links to web_log.conf that appear on the agent UI #6087 (cakrit)
• Fixed ScaleIO collector endpoint paths go.d PR 226 ilyam8
• Fixed web client timeout handling in the go.d plugin httpcheck collector go.d PR 225 ilyam8

Release v1.15.0 contains 11 bug fixes and 30 improvements.

At a glance

We are very happy and proud to be able to include two major improvements in this release: The aggregated node view and the new database engine.

Aggregated node view

The No. 1 request from our community has been a better way to view and manage their Netdata installations, via an aggregated view. The node menu with the simple list of hosts on the agent UI just didn't do it for people with hundreds, or thousands of instances. This release introduces the node view, which uses the power of Netdata Cloud to deliver powerful views of a Netdata-based monitoring infrastructure.

You can read more about Netdata Cloud and the future of netdata here.

New database engine

Historically, Netdata has required a lot of memory for long-term metrics storage. To mitigate this we've been building a new DB engine for several months and will continue improving until it can become the default memory mode for new Netdata installations. The version included in release v1.15.0 already permits longer-term storage of compressed data and we'll continue reducing the required memory in following releases.

Other major additions

We have added support for the AWS Kinesis backend and new collectors for OpenVPN, the Tengine web server, ScaleIO (VxFlex OS), ioping-like latency metrics and Energi Core node instances.

We now have a new, "text-only" chart type, cpu limits for v2 cgroups, docker swarm metrics and improved documentation.

We continued improving the Kubernetes helmchart with liveness probes for slaves, persistence options, a fix for a Cannot allocate memory issue and easy configuration for the kubelet, kube-proxy and coredns collectors.

Finally, we built a process to quickly replace any problematic nightly builds and added more automated CI tests to prevent such builds from being published in the first place.

Acknowledgements

Our heartfelt gratitude for this release goes to the following people:

• @kam1kaze for help with Kubernetes, a fix for the Docker image and documentation improvements.
• @andvgal for the Energi Core daemon collector and the improvement of the python.d plugin.
• @skrzyp1 for improving cgroup monitoring.
• @Daniel15 for the much sought-after "text-only" new chart type.
• @Fohdeesha, @SahAssar, and @smonff for improving the documentation.
• @etienne-napoleone, @karuppiah7890 and @varyumin for their contributions to the Kubernetes helm chart.

Improvements

• Support for aggregate node view #5902 (gmosx)
• Database engine #5282 (mfundul)
• New collector modules:
  • Go.d collectors for OpenVPN, the Tengine web server and ScaleIO (VxFlex OS) instances (ilyam8)
  • Monitor disk access latency like ioping does #5725 (vlvkobal)
• Energi Core daemon monitoring, suits other Bitcoin forks #5894 (andvgal)
• Collector improvements:
  • Add docker swarm manager metrics to the go.d docker_engine collector (ilyam8)
  • Implement unified cgroup cpu limit #5895 (skrzyp1)
  • python.d.plugin: Allow monitoring of HTTP(S) endpoints which require POST data and make the UrlService more flexible #5893 (andvgal)
• Support the AWS Kinesis backend for long-term storage #5914 (vlvkobal)
• Add a new "text-only" chart renderer #5971 (Daniel15)
• Packaging and CI improvements:
  • We can now fix more quickly any problematic published builds via a new manual deployment procedure #5899 (paulkatsoulakis)
  • We added more tests to our nightly builds, to catch more errors before publishing images #5918 (paulkatsoulakis)
• API Improvements:
  • Smarter caching of API calls. Do not cache alarms and info api calls and extend no-cache headers. #5999 (cakrit)
  • Extend the api/v1/info call response with system and collector information #5889 & #5891 (cakrit), #5996 (vlvkobal)
  • k6 script for API load testing #5892 (cakrit)
• Kubernetes helmchart improvements:
  • Added the init container, where sysctl params could be managed, to bypass the Cannot allocate memory issue #18 (kam1kaze)
  • Better startup/shutdown of slaves and reduced memory usage with liveness/readiness probes and default memory mode none #19 (cakrit)
  • Added the option of overriding the default settings for kubelet, kubeproxy and coredns collectors via values.yaml #24 (cakrit)
  • Make the use of persistent volumes optional, add apiVersion to fix linting errors and correct the location of the env field #22, #23 (karuppiah7890)
  • Fix incorrect parameter names in the README #24 (etienne-napoleone)
• Documentation improvements:
  • #5936 (smonff)
  • #5974 (SahAssar)
  • #5992, #6024 (paulkatsoulakis)
  • #6016, #6029, #6030, #6032 (cakrit)
  • #5982 (Fohdeesha)
  • #5980 (kam1kaze)

Bug fixes

• Prowl notifications were not being sent, unless another notification method was also active #6022 (cakrit)
• Fix exception handling in the python.d plugin #5997 (ilyam8)
• The node applications group did not include all node processes. #5962 (jonfairbanks)
• Installation would show incorrect message "FAILED Cannot install netdata init service." in some cases #5947 (paulkatsoulakis)
• The nvidia_smi collector displayed incorrect power usage #5940 (ilyam8)
• The python.d plugin would sometimes hang, because it lacked a connect timeout #5911 (ilyam8)
• The mongodb collector raised errors due to various KeyErrors #5931 (ilyam8)
• The smartd_log collector would show incorrect temperature values #5923 (ilyam8)
• charts.d plugins would fail on docker, when using the timeout command #5938 (paulkatsoulakis)
• Docker image had plugins not executable by user netdata #5917 (paulkatsoulakis)
• Docker image was missing the lsns command, used to match network interfaces to containers #1 (kam1kaze)

Release 1.14 contains 14 bug fixes and 24 improvements.

At a glance

The release introduces major additions to Kubernetes monitoring, with tens of new charts for Kubelet, kube-proxy and coredns metrics, as well as significant improvements to the netdata helm chart.

Two new collectors were added, to monitor Docker hub and Docker engine metrics.

Finally, v1.14 adds support for version 2 cgroups, OpenLDAP over TLS, NVIDIA SMI free and per process memory and configurable syslog facilities.

Acknowledgements

Our contributors kicked the ball out of the park this time. Our thanks go to the following people:
@ekartsonakis for the excellent addition of TLS support to the OpenLDAP collector
@Wing924 whose cat apparently leaves him enough time to help us with springboot2 and a lot more!
@huww98 for his contribution to the NVIDIA SMI plugin.
@varyumin for his help on the Kubernetes helm chart.
@skrzyp1 for the very significant addition of cgroup v2 support
@hsegnitz for his contribution to the web server log plugin.
@archisgore for the quick fixes to the Polyverse-enabled docker image.
@tctovsli for his Rocket Chat notifications improvements.
@JoeWrightss and @vinyasmusic for not letting us get away with spelling mistakes.
@andvgal for the addition to the MongoDB collector.
@piiiggg for the apache proxy documentation fix
@Ferroin for general awesomeness.

Bug Fixes

• Fixed cases where the netdata version produced by the binary or the configure tools of the source code was wrong. Instead of getting something like netdata-v1.14.0-rc0-39a9sf9g we would get a netdata-39a9sf9g. #5860 (paulkatsoulakis)
• Fixed unexpected crashes of the python plugin on macOS, caused by new security changes made in High Sierra. #5838 (ilyam8)
• Fixed problem autodetecting failed jobs in python.d plugin. It now properly restarts jobs that are being rechecked, as soon as they are able to run. #5837 (ilyam8)
• CouchdDB monitoring would stop sometimes with an exception. Fixed the unhandled exception causing the issue. #5833 (ilyam8)
• The netdata api deliberately returned http error 400 when netdata ran in memory mode none. Modified the behavior to return responses, regardless of the memory mode #5819 (cakrit)
• The python.d plugin sometimes does not receive SIGTERM when netdata exits, resulting in zombie processes. Added a heartbeat so that the process can exit on SIGPIPE. #5797 (ilyam8)
• The new SMS Server Tools notifications did not handle errors well, resulting in cryptic error messages. Improved error handling. #5770 (cakrit)
• The installers would crash on some FreeBSD systems, because sha256sum used by the installers is not available on all FreeBSD installations. Modified the installers to properly support FreeBSD. #5760 (paulkatsoulakis)
• Running netdata behind a proxy in FreeBSD did not work, when using UNIX sockets. Added special handling of UNIX sockets for FreeBSD. #5756 (vlvkobal)
• Fixed sporadic build failures of our Docker image, due to dependencies on the Polyverse package ( APK broken state). #5751 (archisgore)
• Fix segmentation fault in streaming, when two dimensions had similar names. #5882 (vlvkobal)
• Kubernetes Helm Chart: Fixed incorrect use of namespaces in ServiceAccount and ClusterRoleBinding RBAC fixes (varyumin).
• Elastic search: The option to enable HTTPS was not included in the config file, giving the erroneous impression that HTTPS was not supported. The option was added. [#5834] (#5834) (ilyam8)
• RocketChat notifications were not being sent properly. Added default recipients for roles in the health alarm notification configuration. #5545 (tctovsli)

Improvements

• go.d.plugin v0.4.0 : Docker Hub and k8s coredns collectors, springboot2 URI filters support.
• go.d.plugin v0.3.1 : Add default job to run k8s_kubelet.conf, k8s_kubeproxy, activemq modules
• go.d.plugin v0.3.0 : Docker engine, kubelet and kub-proxy collectors. x509check module reading certs from file support
• Added unified cgroup support that includes v2 cgroups #5407 (skrzyp1)
• Disk stats: Added preferred disk id pattern, so that users can see the id they prefer, when multiple ids appear for the same device #5779 (vlvkobal)
• NVIDIA SMI: Added memory free and per process memory usage charts to the collector #5796 (huww98)
• OpenLDAP: Added TLS support, to allow monitoring of LDAPS. #5859 (ekartsonakis)
• PHP-FPM: Add health check to raise alarms when the phpfm server is unreachable #5836 (ilyam8)
• PostgreSQL: Our configuration options to connect to a DB did not support all possible option. Added option to connect to a PostreSQL instance by defining a connection string (URI). #5758 (ilyam8)
• python.d.plugin: There was no way to delete obsolete dimensions in charts created by the python.d plugin. The plugin can now delete dimension at runtime. #5795 (ilyam8)
• netdata supports sending its logs to Syslog, but the facility was hard-coded. We now support configurable Syslog facilities in netdata.conf. #5792 (thiagoftsm)
• We encountered sporadic failures of our kickstart installation scripts after nightly releases. We add integrity tests to our pipeline to ensure we prevent faulty scripts from getting deployed. #5778 (paulkatsoulakis)
• Kubernetes Helm Chart improvements: (cakrit) and (varyumin).
  • Added serviceName in statefulset spec to align with the k8s documentation
  • Added preStart command to persist slave machine GUIDs, so that pod deletion/addition during upgrades doesn't lose the slave history.
  • Disabled non-essential master netdata collector plugins to avoid duplicate data
  • Added preStop command to wait for netdata to exit gracefully before removing the container
  • Extended configuration file support to provide more control from the helm command line
  • Added option to disable Role-based access control
  • Added liveness and readiness probes.