LibreQoE/LibreQoS v2.1

on GitHub

Summary

LibreQoS v2.1 is focused on making daily operation easier for Internet Service Providers. The release improves visibility in the Web UI, expands integration compatibility, makes first-run and login behavior clearer, and continues the shaping/runtime work needed for safer reloads and larger deployments.

In practical terms, this release is about helping ISPs spend less time guessing what LibreQoS is doing, less time fighting integration edge cases, and more time using LibreQoS as an operational tool.

Who this release helps most

• WISPs using UISP
• ISPs using Sonar as a source of truth
• operators using Insight and the Node Manager UI heavily
• mixed wireless/fiber ISPs who want to include some subscriber groups and exclude others
• operators onboarding a new box or deploying LibreQoS more broadly

What ISPs will notice

Better visibility in the Web UI

• The Node Manager and executive dashboards have been expanded substantially.
• Operators now get more views into circuit activity, shaped devices, site conditions, queue behavior, CPU usage, and planning-oriented dashboards.
• Bakery, TreeGuard, and StormGuard now have much better UI visibility.

Benefit to ISPs:

• easier to answer "what is happening right now?"
• easier to identify overloaded sites, unhealthy subscribers, or shaping changes
• less need to drop into CLI/debug files for common checks

Improved site map experience

• The Site Map was rebuilt around an OpenStreetMap-based view using Insight-served tiles.
• Map behavior was improved with better viewport fitting, transport scaling, site links, marker sizing, and tile retry/throttling behavior.

Benefit to ISPs:

• easier to orient geographically
• better visual context for site relationships and affected areas
• more useful operational map for day-to-day support and planning

Better first-run and account setup

• First-run setup, login behavior, and related onboarding flows were improved.
• Insight signup and trial handoff flows were also expanded.

Benefit to ISPs:

• cleaner initial setup experience
• less confusion on new installs or reset systems
• easier path into Insight evaluation where desired

Integration improvements

Better Sonar support

The Sonar integration received a major upgrade:

• better GraphQL pagination
• normalized Sonar API URL handling
• better API and non-JSON error messages
• Radius IP support in addition to inventory-item IPs
• child_accounts support
• optional recurring-service fallback mappings for operators whose internet products are modeled outside standard DATA services
• timeout/retry hardening for slower Sonar systems

Benefit to ISPs:

• more subscribers imported correctly
• better compatibility with real-world Sonar deployments
• fewer cases where shaping silently misses customers because Sonar models them differently
• easier troubleshooting when Sonar is slow or returns unexpected responses

Better UISP support

• UISP strategy logic received additional work, including Ethernet advisory support and more validation around edge cases.
• Ignored-only clients are handled more safely in shaping outputs.

Benefit to ISPs:

• more confidence that imported UISP data matches intended shaping behavior
• better handling of subscriber and topology edge cases

Safer handling of ignored subscriber subnets

• Shared integration output handling now does a better job excluding ignored-only subscriber/device rows from generated shaping outputs.
• This was implemented carefully so it does not newly require every imported integration IP to be inside allow_subnets just to survive generation.

Benefit to ISPs:

• easier to exclude whole subscriber populations, such as FTTH customers already shaped at the ONT, without over-billing shaped inventory
• lower risk of upgrade breakage for operators who do not maintain perfect allow_subnets

Operational and shaping improvements

Behind the scenes, v2.1 includes major shaping/runtime work in Bakery and TreeGuard.

• substantial Bakery diff/apply improvements
• queue math and qdisc handle work
• more large-scale and integration-oriented validation
• continued TreeGuard runtime control and state handling work
• stronger node and circuit identity handling

Benefit to ISPs:

• safer live refreshes
• better behavior at larger scale
• lower chance of shaping state drifting or becoming harder to reason about during operational changes

Configuration and override improvements

• queue_mode is now more visible in the config/runtime model, with clearer shape vs observe behavior.
• Operator override handling and stable node identity handling were improved.
• Network JSON support was extended in ways that better support stable node identifiers and transport metadata.
• CPU topology/default handling and queue tracking were refined further.

Benefit to ISPs:

• easier to move between observe-only and shaping workflows
• more predictable overrides
• better long-term maintainability for integration-driven deployments

Upgrade notes

• If you use Sonar, review import counts after upgrading. This release improves support for Radius IPs, child accounts, and recurring-plan fallback behavior.
• If you use UISP, verify expected subscriber/device import behavior after upgrade as usual, especially on networks with unusual topology or mixed address handling.
• If you use the Site Map, note that the newer map behavior depends on the Insight-served tile path used by the UI.
• If you use observe mode, review queue_mode after upgrade and confirm the box is in the intended mode.
• If you are onboarding a new box or restoring one from a minimal state, expect the first-run and login flow to behave differently than older releases.