Netdata

Netdata v2.10.4 is a patch release focusing on stability, memory safety, and crash resilience, featuring new macOS hardware sensor collectors and various database engine improvements.

Netdata v2.10.3 patches EBPF per-PID memory leak, switches SNMP uptime to snmpEngineTime, allows dots in dynamic config names for domains, and removes unused extra_details from go.d/powerstore to fix decoding errors.

Netdata v2.10.2 patches ZFS-related diskspace crashes, reduces SNMP MaxOIDs to 20 and removes 32-bit counter fallbacks, and eliminates false-positive timeout warnings by removing the wait-decision timeout and hardening the dyncfg path.

Netdata v2.10.1 patches SNMP with ping_only mode, fixes retry serialization and bulk-walk logic, and reduces 503s by buffering the go.d command channel and extending handoff timeouts.

Netdata v2.10.0 adds secrets management, Nagios plugins collector, Azure Monitor support, AI Conversations, enhanced alerting and dashboards, broader collector coverage, and OpenTelemetry improvements for stability.

Netdata v2.9.0 adds interactive cross-database query analysis, a Go-based MSSQL collector, OpenTelemetry log ingestion with journal-backed storage (beta), SNMP network interfaces, and overall stability improvements.

Netdata v2.8.4 is a patch release that adds Go version in build info and updates the Go toolchain to v1.25.5.

Netdata v2.8.3 is a patch release fixing go.d station handling, Windows path detection, non-blocking job shutdown, Windows AD/ADCS/ADFS metrics, SNMP discovery, MSSQL, MongoDB, RabbitMQ compatibility, ndexec output, and updated components/packaging.

Netdata v2.8.2 patches multiple issues: Windows sensors init moved to the sensor thread; OTEL metric names use slashes for hierarchy; Proxmox group assignment fixed; NVIDIA device access granted; replication loops/stalls fixed; runtime config dir precedence; Redis logs silenced.

Netdata v2.8.1 patches several issues: persists vnodes after restart, disables PSS memory sampling by default, adds metric unit source to alerts, reduces Windows datafile sync logs, fixes compile when dbengine is disabled, and improves Podman container ID extraction in the cgroups plugin.

Netdata v2.8.0 adds SOC 2 Type 2 compliance, Expanded Chart Analysis, AI-log visualization, scheduled Insights/Investigations reports, ServiceNow integration, alpha generic SQL collector, PSS memory estimation, plus stability fixes and SNMP legacy collection removal.

Netdata v2.7.2 is a patch release addressing Windows crash deadlocks, startup/log handling, disk space and memory reporting fixes, and numerous improvements including SNMP, OpenTelemetry host ID fallback, installer reliability, and remote MCP support.

Netdata v2.7.1 patches issues since v2.7.0: adds support for the old RabbitMQ whoami tags format and fixes Windows Agent that prevented go.d plugin from starting.

Netdata v2.7.0 adds Netdata AI, chart annotations, quick data export, OpenTelemetry alpha ingestion, SNMP profiles as default/stable, default IP detection across platforms, plus broader stability fixes and SNMP legacy collection deprecated by default.

Netdata v2.6.3 patches issues since v2.6.2: adds a build-time check for bad compiler flags; fixes saving of virtual host labels and the live flag on restart; and enhances SNMP profiles and the go.d/snmp collector.

Netdata v2.6.2 is a patch release addressing PUBACK handling and 32-bit timeout corruption, adds extensive ACLK telemetry and latency metrics, fixes virtual node inheritance, security validations, and updates OpenSSL/curl with broader CI support.

Netdata v2.6.1 patches multiple fixes: reset release state, dynamic thread init, move worker completion signaling, improve SNMP metadata/UCUM, query NVMe at controller level, fix build timestamps, ACLK MQTT timeout/buffer, eBPF CO-RE 1.6.0, and SSL cert paths on Rocky Linux/static builds.

Netdata v2.6 adds MCP server integration and AI Insights with natural language queries, Okta SCIM provisioning, alpha SNMP profile-based monitoring, native RHEL/AlmaLinux/Rocky 10 packages, and a Rust systemd-journal plugin for static builds.

Netdata v2.5.4 patches multiple fixes and optimizations, including Go plugin label sanitization and faster startup, Windows -NoProfile, reduced memory footprint via ARAL allocator, 32-bit build fixes, TLS for rediss, 429 exclusion in weblog, registry save backoff, faster shutdown, safer query progress, and vendored Protobuf/Abseil for static builds.

Netdata v2.5.3 patches multiple issues: fixes hub queue context, enables debugging for go.d.plugin jobs and HTTP checks, resolves duplicate names in dynamic config, corrects Oracle tablespace usage, and improves file rotation/indexing with better scheduling.

Netdata v2.5.2 patches crashes and memory issues, updates eBPF to 1.5.1, fixes virtual node/config, metadata/SNMP/maintenance/logging issues, division-by-zero protection, MSSQL threading, and related minor fixes.

Netdata v2.5.1 patches multiple fixes including obsolete chart cleanup, SNMP counters/descriptions improvements, SNMPv3 credential naming, uv_sleep removal for older libuv, Docker claim handling, optional thread stack size, and CouchDB size unit fix.

Netdata v2.5.0 prioritizes stability with enhanced crash/memory/thread handling, deadlock detection, robust resource cleanup, and protection against corrupted data, plus broader platform support and refined packaging/documentation.

Netdata v2.4.0 emphasizes stability with numerous crash/memory-leak fixes, broad memory optimization (SQLite, containers, DB engines), and introduces optional Windows Service Monitoring (windows.plugin/PerflibServices).

Netdata v2.3.2 is a patch release fixing numerous issues since v2.3.1, including journal/file handling, memory safety, shutdown crashes, UUID generation, telemetry suppression, Windows/ACLK fixes, and performance/diagnostics enhancements.

Netdata v2.3.1 is a patch release fixing build/config issues, memory monitoring improvements, security hardening, crash diagnostics enhancements, and added system metrics, plus updated updater paths and ACLK/DB handling for more stable operation.

Netdata 2.3 introduces ephemeral node handling to reduce alert noise, extreme cardinality protection, zero-sampling crash reporting, and automatic SNMP discovery, boosting reliability, scalability, and ease of monitoring.

Netdata v2.2.5 patches multiple fixes including MariaDB CPU-time, replicated-readonly clarity, offline-agent claim handling, stats.d median/percentile interpolation, startup/shutdown DB safety, memory/dequeuing, journal watcher, NVMe path handling, MySQL globals refresh, and CI/ARM build improvements.

Netdata v2.2.4 patch fixes go.d cleanup of stopped jobs, fixes health memory leak, adds agent status for Cloud connection in progress, and fixes Windows MSSQL instance monitoring.

Netdata v2.2.3 is a patch release fixing: newly created/claimed agents not appearing in cloud, database engine 32-bit limits for files >4GB, and stale contexts after child reconnections due to cached collection flags.

Netdata v2.2.2 patch fixes memory corruption in streaming, MSSQL named instance detection, receiver-sender socket errors, simplifies eBPF dependencies, removes lm_sensors from Go RPM, fixes proxy credential decoding, adjusts alert timing, enhances buildinfo, and updates NVMe discovery for nvme-cli v2.11.

Netdata v2.2.1 is a patch release fixing negative metric display, performance/SQL optimizations, removed obsolete ZFS option, enhanced streaming, alerts, CPU, memory, and various crash/health fixes; includes network plugin tweaks and ping improvements.

Netdata 2.2.0 reduces memory by 50% across deployments, halves bandwidth for parent-to-parent streaming, cuts CPU usage by 20%, adds auto-detected profiles (Parent/Child/IoT), and introduces a single-node dashboard plus customizable space navigation.

Netdata v2.1.1 patches several issues: fixes health subsystem file descriptor leaks, updater.sh temp-dir handling and cleanup, RabbitMQ API permission handling, duplicate Prometheus metadata, and an AMDGPU DRM data/crash issue.

This release centers on streaming scalability with CPU and memory optimizations, smarter query handling, plus Cloud features (auto room assignment, alert repeats, dashboard favorites) and bulk dynamic configuration for collectors and alerts.

Netdata v2.0.3 patches issues since v2.0.2 and sets Windows service startup to Automatic (Delayed Start) to improve reliability after reboots; reinstall should apply this automatically, with manual steps if needed.

Netdata v2.0.1 patches exit crashes (replication), fixes Docker deadlock from spinlock, adds HTTP caching headers, fixes go.d/dyncfg dynamic config validation, and excludes unmounted volumes from Windows disk alerts.

Netdata 2.0 introduces native Windows support,_UI3 as a separate package_, API v3, enhanced process/network monitoring, SSO enhancements, a referral program, UTF-8/intuive units, and deprecates v0–v2 dashboards and older APIs.

Netdata v1.47.4 patches journald messaging, updates OpenSSL 3.3.2 and curl 8.10.0, fixes CI cross-compilation for go.d, adds runtime swap-change checks, corrects LXC disk/RAM calculations, fixes the nvidia_smi loop timeout, and adds 64-bit atomic support checks (RISCV32).

Netdata v1.47.3 patches: fix smartctl exit-status evaluation, Postgres 17 compatibility, uwsgi metric write-order issue, and updater.sh replacing /dev/stdout symlink.

Netdata v1.47.2 is a patch release fixing issues since v1.47.1, including reduced connection timeout and IPv4 fallback for ACLK connections.

Netdata v1.47.1 patch fixes streaming receivers crash, improves alert handling (pending/removal after restart), fixes drive data parsing, gorilla compression buffer overflow, supports symlinked vnode configs, prevents stale node reappearance, clarifies netdatacli help, and fixes DHCP parse panic.

Netdata Agent 1.x ends with 1.47; key updates include SNMP automatic interface discovery, Python-to-Go plugin migrations, SCIM identity integrations, enhanced reachability notifications, IPv6 support, stale-node removal, and MS Teams workflow changes.

Netdata v1.46.3 patches missing dashboard contexts on fresh installs, fixes the "Directory existence" chart label, and corrects nsudo execution by forcing uid/gid/egid to 0 before locating the executable.

Netdata v1.46.2 patches several fixes: set root UID/GID, gate ksm/swap/cma/zswap metrics behind features, align Go version, add lm_sensors weak dep, fix Prometheus export, switch IO times to microseconds, improve Docker host handling, and enable updater service.

Netdata v1.46.1 patches several fixes: Rspamd collector handles responses without FuzzyHashes; validates dyncfg and plugin job names (no invalid chars or colons); fixes cgroups v1 detection on non-systemd nodes; and corrects space percentage in the retention chart.

GA Dynamic Configuration with in-dashboard alert/collector setup and multi-node deployment, Native Windows support in progress, AWS Marketplace subscriptions, time- and space-based retention with new charts, ZSTD dbengine compression, ClickHouse collector, UCUM-ready units, and expanded notification integrations.

Netdata v1.45.6 is a patch release fixing service discovery for apps with no command-line params (go.d.plugin), host label updates in Cloud on label reload, and ACLK Proxy MQTT CONNECT Host header handling.

Netdata v1.45.5 patches multiple issues: fixes streaming payload corruption, crashes from missing dimension IDs, ACLK proxy Host header, adds ML support to CentOS 7 RPM, corrects cockroachdb collector calc, enables offline updater, ensures Cloud Alerts emit REMOVED for disconnected children, and adds vnode support in go.d/windows.

Netdata v1.45.3 patches a local privilege escalation in 1.45.x, urges upgrades for 1.45.0–1.45.2, and fixes ndsudo path restrictions, alert percentage lookups, macOS LaunchDaemons uninstall, and increases the default thread stack size to 1 MB.

Netdata v1.45.2 patches several issues: auto-discovery for SQL listeners (passwordless Unix sockets), negative host/chart label matching in alerts, go.d.plugin stability and TLS Unbound panic fixes, HTTP UserAgent, enhanced Nginx status endpoint discovery, libyaml linking fix, and stricter go.d config validation.

Netdata v1.45.1 fixes config defaults, enhances go.d.plugin discovery (IPv6, Apache), improves OpenSSL/macOS, removes USR1 save references, adds optional ZSTD support, fixes reference counts, enables Gorilla compression, and tightens tests/builds.

Netdata 1.45 adds a mobile app, drag-and-drop dashboards, network viewer, dynamic configuration, macOS process monitoring, a Homelab plan, a new CMake/Ninja build system, and moves the Go plugin into the main repo.

Netdata v1.44.3 patches slow query/ACLKSYNC CPU handling, enables graceful exit on listen bind failures, fixes alert durations and analytics setup race, and updates the base image with Debian backports.

Netdata v1.44.2 patch fixes logging consistency and container detection, stabilizes multiple plugins (cgroups, chartd.d, cups), improves macOS/FreeBSD handling, prevents crashes, and updates go.d.plugin to v0.58.0.

Netdata v1.44.1 fixes uninstall, SIGPIPE, memory leaks, build-info values, single-core streaming, CentOS7/Amazon Linux 2 packaging, AMD memory clock crash, Debian postinst, home-dir bug, and adds a lightweight editor to the Docker image.

Netdata v1.44 improves performance vs Prometheus, enables systemd-journal logging with UI, ships log2journal, expands Functions, and adds alert destinations (Amazon SNS and Telegram) for Netdata Cloud.

Netdata v1.43.2 patch fixes rrdlabels type, label copy with new keys, label propagation for streaming metrics, faster Cloud alert queries on large-parent deployments, systemd-journal field filtering, and proper propagation of updated child labels to parents.

Netdata v1.43.1 fixes build issues, pattern matching, query planning for time periods, Discord payloads, discovery deadlocks, duplicate labels, dimension checks, 32-bit dbengine stats, and shutdown improvements, plus go.d updated to v0.56.3.

Major update with a nearly rewritten systemd-journal featuring histograms, live play, field filtering and colorization; improved VMware vSphere monitoring; perf/alerts/ui improvements; lower resource use; 32-bit Netdata on 64-bit hosts; on‑prem Netdata Cloud.

Netdata v1.42.4 is a patch release addressing a bug where alarm variables weren’t created for all chart dimensions, improving reliability since v1.42.3.

Netdata v1.42.3 patches a Prometheus exporter memory leak and fixes handling of closed connections in streaming.

Netdata v1.42.2 patches numerous issues: fixes to plugin dependencies, Prometheus label dedup, RAM calculation, static OpenSSL (32-bit/threading), stable release channel, multi-collector chart crashes,

Netdata v1.42.1 is a patch release fixing Systemd-journal/Processes entries, Docker stdout, Health API, proxy 401s, packaging/deps, journaling, retention, drives, systemd-journal restart, OpenSUSE packaging, and updating UI to v6.28.0 and OpenSSL to 1.1.1v.

Netdata v1.42.0 adds a beta Integrations Marketplace with 800+ integrations, a new systemd-journal query function, UI-based agent claiming to Netdata Cloud, and an anomaly-scoring badge for quick dashboard anomaly insight.

v1.41 introduces a unified Agent Dashboard with ML-first dashboards and multi-node support, plus Netdata Assistant, a FreeIPMI collector for enterprise hardware, and FD leak detection in apps.

Netdata v1.40.1 fixes ebpf crashes and slow Termination, improves building on RPM/macOS/OpenSUSE, resolves health log crash, SSL retry handling, plugin ownership in static builds, Nvidia SMI exception, wrong systemd unit, virtual NIC charts, and i386/native package arch detection.

This release adds dashboard summary tiles across many views, introduces Cloud alert silencing rules, extends ML training to 24 hours with persisted multi-model per metric, and rewrites SSL support for the Netdata Agent.

Patch fixes auto-update permissions for static installs; improves recipient handling in health_alarm_notify.conf; fixes InfiniBand bandwidth collection; prevents memory buffer overflows; fixes two ACLK memory leaks; suppresses noisy metrics in ML to prevent wrong models; improves ML transaction rollback.

Netdata Charts v3.0 introduces NIDL-enabled, anomaly-aware charting in Cloud, adds Windows monitoring with 170+ metrics, supports virtual nodes and custom labels for large deployments, and previews packaging, zlib dependency, and upgrade changes for future releases.

Patch release for v1.38 upgrades OpenSSL to v1.1.1t in static and Docker builds to fix vulnerabilities, plus fixes for empty-data dimensions, shutdown crashes without eBPF, older systemd handling, and BSD/Linux kernel build issues.

Netdata v1.38.0 ships a new dbengine v2 with lower memory and huge performance gains, adds runtime functions, events feed, expanded Cloud notifications (Slack, PagerDuty, Discord, Webhooks), richer RBAC, and broad improvements across integrations, ML, visuals, packaging, and deprecations.

Netdata v1.37.1 patches crashes in parent/children connections, dbengine rotation, and metrics alignment; fixes Fedora packaging and backfilling statistics; and adds performance optimizations for workers, replication, and dictionaries.

Netdata v1.37 fixes two security issues (streaming authorization and alarm-notify command execution) while delivering infinite scalability, default multi-tier DB retention, enhanced Grafana data source, PostgreSQL monitoring rewrite, and overall performance improvements.

Netdata v1.36.1 patches two issues: 32-bit agents crashing during cloud data exchange, and a Go plugin installer bug that blocked the new WireGuard collector without user intervention.

Release v1.36 adds a fast Volume metric correlation algorithm, MC/Anomaly Advisor integration with a new dashboard, Tiering for long-term metrics, Kubernetes state support, anomaly rates on all charts, a centralized admin interface with bulk deletion, metadata syncing, visualization tweaks, and new chart labeling.

Netdata v1.35.1 is a patch that fixes a static-build install bug that accidentally disables automatic updates during updates (affecting static installs since 2022-03-22 unless updates were disabled); see GitHub issue 13102 for re-enabling.

This release adds on-device ML Anomaly Advisor, instant agent-side metric correlation, Kubernetes monitoring improvements, visualization/alerts/nodes enhancements, StatsD with dictionaries, 3x faster agent queries, streaming refinements, broader optimizations, and a tech-preview MQTT 5 client.

This patch fixes versioning errors in Netdata v1.34.0 release: incorrect version strings on artifacts and binaries, Docker image tag inconsistencies, and partially duplicated Git tags.

Kubernetes CPU throttling charts added; ML anomaly detection optimized; streaming compression enabled by default (Alpha); SNMP collector rewritten in Go; improved installation via kickstart enhancements.

Netdata v1.33.1 is a patch release with bug fixes and documentation updates, plus a Cloud migration banner for nodes on the old architecture.

Release v1.33 adds pre-built Linux packages, streaming compression (tech preview), and eBPF CO-RE support, while deprecating legacy features and updating the kickstart installer for smoother, package-based installs.

Netdata v1.32.1 is a patch release fixing bugs (cloud context, swap alarm, AlarmLogHealth init), updating ACLK/ML documentation, and applying packaging tweaks.

v1.32.0 introduces a new Cloud backend with incremental agent-cloud protocol and parent/child support, expanded eBPF latency monitoring, opt-in ML anomaly detection, a timezone/time controls UI, POWER8+ Docker support, plus numerous collectors and fixes.

Netdata v1.31.0 introduces a redesigned dashboard, enhanced Linux kernel insights via eBPF, on-node ML with Z-scores and ChangeFinder, plus 10 new collectors, 54 improvements, 31 docs, and 29 bug fixes.

Patch release addressing issues since 1.30.0: minor documentation fixes, packaging improvements including OpenSSL 1.1.1k and dist-tarball fixes, plus bug fixes for memory corruption, shutdown crash event, and health log entries.

Netdata v1.30.0 overhauls packaging, replaces Google Analytics/GTM telemetry with PostHog, introduces ACLK-NG and a developer devenv, adds 3 new collectors plus 38 improvements (13 in the dashboard), 16 docs, 17 bug fixes, and initial UI/alarms updates.

Release v1.29.3 patches Netdata Agent stability by fixing proc.plugin crash on obsoleted charts (RRDSETVAR), plus packaging tweak for static LWS in RPMs and several docs updates and a general bugfix.

Patch release v1.29.2 fixes a bug that set all os_* host labels to unknown, adds Opsgenie API URL configurability, and includes several improvements and multiple bug fixes.

Release v1.29.1 hotfix fixes crash during shutdown caused by a locking bug in the cgroups internal plugin, preventing excessively long exit times of the Netdata Agent.

Hotfix v1.28.0 fixes a deadlock by correcting locking after on_connect failure in ACLK, ensuring the web server lock is released properly.

v1.27.0 adds SQLite-based metadata storage with a faster, lighter startup, a new time/date picker, an extent cache, two new collectors (anomalies and alarms), plus 54 improvements, 41 docs, and 58 bug fixes.

Release v1.26 adds three Go-based collectors (filecheck, systemdunits, isc_dhcpd), StackPulse notification integration, a redesigned Netdata Learn docs, and broader Cloud-ready compatibility, plus 21 improvements, 13 docs updates, and 12 bug fixes.

This release adds an Elasticsearch collector, improves Prometheus filtering/grouping and ACLK performance, and includes 59 bug fixes, 27 improvements, and 15 documentation updates across the stack.

Adds a generic Prometheus/OpenMetrics collector and a new multi-host database mode enabling distributed queries and shared resources, with backward-compatible legacy paths.

Netdata v1.23.2 patches a JSON parsing buffer overrun vulnerability and delivers assorted bug fixes and improvements (including Infiniband monitoring, packaging tweaks, and CI adjustments).

Netdata v1.23.1 patches segfaults from simultaneous obsolete/archived dimensions, fixes duplicate alarm IDs, and includes multiple additional bug fixes (Kubelet cgroups, stream, registry, virtualization, unittest, and more).

v1.23.0 adds Kubernetes service discovery (22 services), enhanced eBPF metrics at event frequency, a new exporting engine for multiple backends, persistent metadata, and broad improvements and bug fixes.