Release v1.25.0

The v1.25.0 release of the Netdata Agent is focused on improving Netdata's usability across the board. We added more customization to how the Prometheus collector implemented in v1.24 meaningfully visualizes metrics. In addition, we've focused on fixing bugs and ensuring that core functionality of the Netdata Agent, such as the ACLK, works more efficiently.

This release contains 1 new collector, 27 improvements, 15 documentation updates, and 59 bug fixes.

At a glance

Improved filtering and grouping for the Prometheus collector gives you more flexibility in how Netdata collects and visualizes metrics from more than 600 Prometheus endpoints. The Prometheus collector is designed to visualize every metric exposed on a Prometheus endpoint generically, but one chart for every metric is often not the most meaningful presentation. Filtering and grouping options bring the same "bespoke" feeling that you find in our other collectors, such as having input/output metrics on a single chart instead of two.

You can read about filtering and grouping in our documentation right now. If you haven't heard about the generic Prometheus collector, read our v1.24 blog post for details on why we continuously make Netdata more interoperable with other monitoring solutions.

We also made significant improvements to the robustness and responsiveness of the Agent-Cloud link (ACLK), which is used to stream metrics and alarm status if you sign up for Netdata Cloud and claim your nodes. The disconnect and reconnect process is now more reliable, and all metrics data is now Gzip compressed. Now that the payloads are smaller and more quickly processed, you'll see improved responsiveness when viewing dashboards in Netdata Cloud.

We added a new Elasticsearch collector, written in Go, to help you collect metrics from and monitor Elasticsearch instances. This collector is preinstalled with the Netdata Agent and often works with zero configuration, but can also be tweaked to collect only specific stats, gather metrics with TLS, and more. See the documentation for generated charts and configuration options.

Acknowledgments

We're grateful to the Netdata community for a huge wave of contributions for this release.

• @mklepaczewski for adding a JSON log parser to the go.d/web_log collector.
• @glesys-andreas for adding socket support for the go.d/phpfpm collector.
• @K900 for adding and documenting how to read container names from Podman.
• @pando85 for fixing the link to Caddyfile's basicauth in the Docker documentation.
• @roedie for improving Debian packaging by streamlining control and init files.
• @vsc55 for adding support for IP ranges in the python.d/isc_dhcpd collector.
• @mrbarletta for fixing MySQL collector documentation to mention the netdata user.
• @Saruspete for fixing RPM build script version issues.
• @michmach for improving the uninstall script to correctly state if the group was deleted.
• @Steve8291 for removing PrivateMounts in systemd journal logs.
• @mrbrutti for updating netdata-installer.sh to enable Netdata Cloud support in macOS.
• @weijing24 for adding RAM info for macOS to system-info.sh.
• @scottymuse for fixing latency-avg chart units in the python.d/dnsdist collector.
• @Ancairon for improving proc.plugin to collect the active processes limit on Linux systems.
• @scatenag for fixing TLS over LDAP in the python.d/openldap collector.
• @florianmagnin for adding new options to the python.d/varnish collector for multiple storage backends.
• @devinrsmith for fixing the print message when building for Ubuntu Focal.

Improvements

• Add code to release memory used by the global GUID map (#9729, @stelfrag)
• Add check for spurious wakeups (#9751, @vlvkobal)

Netdata Cloud

• Add v2 HTTP message with compression to ACLK (#9895, @underhood)
• Add version negotiation to ACLK (#9819, @underhood)
• Add claimed_id for child nodes streamed to their parents (#9804, @underhood)
• Update netdata-installer.sh to enable Netdata Cloud support in macOS (#9360, @mrbrutti)

Collectors

• Update go.d.plugin version to v0.22.0 (#9898, @ilyam8)
  • Add JSON parser to weblog collector (#417, @mklepaczewski)
• Update go.d.plugin version to v0.21.0 (#9881, @ilyam8)
  • Add new Elasticsearch collector (#421, @ilyam8)
  • Add filtering option to Prometheus collector (#416, @ilyam8)
  • Add custom grouping option to Prometheus collector (#418, @ilyam8)
  • Add add socket support to PHP-FPM collector (#402, @glesys-andreas)
• Add support for IP ranges to Python-based isc_dhcpd collector (#9755, @vsc55)
• Add Network viewer charts to ebpf.plugin (#9591, @thiagoftsm)
• Add collecting active processes limit on Linux systems (#9843, @Ancairon)
• Improve eBPF plugin by removing unnecessary debug messages (#9754, @thiagoftsm)
• Add CAP_SYS_CHROOT for netdata service to read LXD network interfaces (#9726, @vlvkobal)
• Add collecting maxmemory to python.d/redis (#9767, @ilyam8)
• Add option for multiple storage backends in python.d/varnish (#9668, @florianmagnin)

Dashboard

• Update dashboard v1.4.2 (#9837, @jacekkolasa)
  • Disable calls to netdata.cloud when --disable-cloud option is used during installation (#114, @jacekkolasa)
  • Fix Y-axis and auto-scaling for constant values (#115 & (#117, @jacekkolasa)
  • Fix broken dashboard when browser is configured to have no preferred language (#118, @jacekkolasa)
  • Fix d3-pie chart unit conversion on updates (#119, @jacekkolasa)
• Update dashboard to v1.3.1 (#9786, @jacekkolasa)
  • Fix stacked chart dimension visibility (#113, @jacekkolasa)

Packaging/installation

• Improve handling of offline installs (#9805, @Ferroin)
• Improve Debian packaging by streamlining control and init files (#8982, @roedie)
• Remove dependency on libJudy for systems which don't have it (#9859, @Ferroin)
• Add code to bundle libJudy on systems which do not provide a usable copy of it (#9776, @Ferroin)
• Improve temporary directory checking in installer and updater (#9797, @Ferroin)
• Add proper certificate handling cURL in our static build (#9733, @Ferroin)

Documentation

• Improve and correct vulnerability reporting instructions (#9696, @cakrit)
• Fix broken link in privacy policy (#9771, @joelhans)
• Update supported collectors doc to organize by type (#9513, @joelhans)
• Change instruction to reload HEALTH (#9869, @thiagoftsm)
• Fix typo in health documentation (#9860, @thiagoftsm)
• Fix broken Edit this page link in simple patterns doc (#9847, @joelhans)
• Remove Google Charts info from API doc (#9826, @joelhans)
• Fix broken link and clean up frontmatter in health docs (#9813, @joelhans)
• Improve dbengine docs and add new multihost setting (#9817, @joelhans)
• Improve health docs by adding daemon config to health section and standardizing IP references (#8837, @joelhans)
• Add and document support for reading container names from Podman in cgroups.plugin (#9474, @K900)
• Fix docker packaging caddyserver basicauth link (#9812, @pando85)
• Fix MySQL collector documentation to mention netdata user (#9555, @mrbarletta)
• Add community link to readme (#9602, @zack-shoylev)
• Add v1.24 news to main README (#9721, @aabatangle)

Bug fixes

• Fix setting the default value of the home directory to the environment's HOME (#9711, @cakrit)
• Fix memory mode none not dropping stale dimension data (#9917, @mfundul)
• Fix memory mode none not marking dimensions as obsolete (#9912, @mfundul)
• Fix race condition with orphan hosts (#9862, @mfundul)
• Fix the log level in cgroup-network helper (#9836, @vlvkobal)
• Fix empty dbengine files (#9820, @mfundul)
• Fix timestamps for global variables in Prometheus output (#9779, @vlvkobal)
• Fix long stats.d chart names (suggested by @vince-lessbits) (#9783, @amoss)
• Fix HTTP header for the remote write exporting connector (#9775, @vlvkobal)
• Fix netfilter to close when receiving a SIGPIPE (#9756, @thiagoftsm)
• Fix exporting update point (#9748, @vlvkobal)
• Fix for ignored LXC containers (#9645, @vlvkobal)
• Fix issue with missing alarms (#9712, @stelfrag)
• Fix child memory corruption by removing broken optimization in the sender thread (#9703, @amoss)
• Fix crash when receiving malformed labels via streaming. (#9715, @mfundul)
• Fix collectors on MacOS and FreeBSD to ignore archived charts. (#9695, @mfundul)
• Fix sending follow-up alarms when the initial status matches the notification (#9698, @cakrit)
• Fix typo in option name used to use bundled libJudy (#9893, @prologic)
• Fix handling of libJudy bundling for RPM packages (#9875, @Ferroin)
• Fix another typo in the libJudy bundling code (#9904, @Ferroin)
• Fix missing newline concatentation slash causing failures in RPM builds (#9900, @prologic)
• Fix high CPU in IPFS collector by disabling call to the /api/v0/stats/repo endpoint by default (#9687, @ilyam8)
• Fix flushing errors (#9738, @mfundul)
• Fix bugs in handling of Python 3 dependencies on install (#9839, @Ferroin)
• Fix RPM build script version issues (#9808, @Saruspete)
• Fix installation to not install eBPF plugin components when they shouldn't be installed (#9844, @vlvkobal)
• Fixed tmpdir handling failure on macOS/FreeBSD. (#9842, @Ferroin)
• Fix netdata-uninstaller.sh to correctly state whether the group was deleted (#9835, @michmach)
• Fix updater bug introduced by incomplete variable rename in #8808 (#9834, @Ferroin)
• Fixed bug in installer introduced by #8808 (#9831, @Ferroin)
• Fix systemd journal logs to remove PrivateMounts (#9619, @Steve8291)
• Fix netdata-updater.sh to correctly pass REINSTALL_OPTIONS (#8808, @prologic)
• Fix handling of offline installs (#9805, @Ferroin)
• Fix install if system does not have ebpf.plugin (#9809, @roedie)
• Fix packaging to enable eBPF collector only if enabled in config.h (#9752, @Saruspete)
• Fix numerous bugs in duplicate install handling (#9769, @Ferroin)
• Fix netdata/netdata Docker image size (#9669, @prologic)
• Fix global GUID map memory leak (#9725, @stelfrag)
• Fix buffer overflow in rrdr structure (#9903, @mfundul)
• Fix HTTP error messages in alarm notifications (#9887, @thiagoftsm)
• Fix lock order reversal (Coverity defect CID 361629) (#9888, @mfundul)
• Fix missing macOS RAM info in system-info.sh (#9882, @weijing24)
• Fix latency-avg chart units in python.d/dnsdist (#9871, @scottymuse)
• Fix TLS over LDAP in the python.d/openldap collector (#9853, @scatenag)
• Fix multi-host DB corruption when legacy metrics reside in localhost. (#9855, @mfundul)
• Fix compilation warnings on FreeBSD (#9845, @underhood)
• Fix proxy forwarding claim_id to old parent (#9828, @underhood)
• Fix old dashboard third-party packaging (#9814, @jacekkolasa)
• Fix loading custom dashboard_info in /old dashboard (#9792, @jacekkolasa)
• Fix unit tests for exporting engine (#9766, @vlvkobal)
• Fix code formatting for the mdstat collector (#9749, @vlvkobal)
• Fix health notifications configuration to clarify which notifications are received when the "|critical" limit is set (#9740, @cakrit)
• Fix print message when building for Ubuntu Focal (#9694, @devinrsmith)
• Fix alarm redirection link for Cloud to stop showing 404 (#9688, @cakrit)