Dwarfguard version 2.0.2 "Keeper patch 2"
- status: Actual
- released at: 2026-06-28
- supported modes of operation:
- production: Yes
- free usage: Yes
- 2.0.2 build 0
- SHA-256 sum: af6e29be6c404ca4657e613fef7f44d7a8463f8b72ab0b2bcc207707e10cbe03
- supported deployment modes:
- DT SAAS / DT Cloud / DT Server Managed: Yes / Yes / Yes
- Partner SAAS / Partner Cloud / Partner Server Managed: Yes / Yes / Yes
- Customer self-managed: Yes
- Private network: Yes
- support planned: major release, support via patch releases 2026+
- upgrade supported from:
- target audience: All
- showcase video: Dwarfguard 2.0 showcase video
- paper: Dwarfguard 2.0.0 paper
Release notes - known issues, changes since last release
- Version 2.0.2 is a first patch on the stable 2.0 branch.
- This patchlevel fixes issues in the DCL layer that were introduced in the 2.0.0 version and partially reported for 2.0.1.
- This patchlevel does NOT fix all the known issues for Dwarfguard 2.0.1. It was deemed necessary to release the DCL changes as these are affecting our customers while leaving fixing the lower-impact issues for later releases
Featured changes
- Fixes for following reported issues:
- Changes Tracker is too slow responding when there are many data entries
- Grafana integration command (
dwarfg_grafint) does not work for Grafana version 13.x
- Further details:
- This patchlevel also updates the DCL behavior for the SQL data export (there are 3 types of DCL export, only the SQL DB is updated)
- In Dwarfguard 1.x, the DCL export for SQL logged only changes in the selected metrics (like the device temperature change). In 2.x, all reported data were stored in DCL SQL export. The reason was for the Changes Tracker to have data for every device when data window is limited. However, it introduced a performance bottleneck for customer running Dwarfguard on VM with limited I/O performance managing hundreds of devices, potentially making the Changes Tracker very slow.
- Patchlevel 2.0.2 (and also the upcoming 2.1.0) introduces new behavior - there are two situations when the metric is logged:
- Reported value has changed - this means every data change for the selected metrics is logged
- Last reported metric is older than the set interval (24 hours since service start by default) - this means that there are data for the Changes Tracker even when the time window is limited (e.g. when you want to see the highest reported temperature between your devices in last 5 days)
- Grafana integration command
dwarfg_grafintwas fixed so that it now supports Grafana 13.
- Both Debian GNU/Linux 12 and Debian GNU/Linux 13 are supported as underlaying OS
Known issues (updates with bugs being reported)
- "Download preset agent" button in Device details leads to error 500 for OpenWRT devices
- Reported: 2026-04-11 for version 2.0.0
- Impact: None (obsolete functionality handled otherwise in 2.0.0)
- Severity: None
- Status: Unfixed. The button will be removed in 2.1.0
- Details: In historic Dwarfguard versions (0.7.x and 0.8.x), a specific device-preset agent was needed to properly reinstall the agent on a device so that this would not loose its Device ID. Nowadays, this is handled by Machine ID, which is generated automatically and stays the same for the same device even on agent reinstall. This functionality is obsolete and should have been removed even in version 1.0.0 however it stayed there as an left over. There is no point in using this functionality.
- Workaround: Do not click on the button as it does not provide anything of value.
- When deleting a device, GUI shows "Device was not deleted" (while it was)
- Reported: 2026-05-27 for version 2.0.1
- Impact: Cosmetic/confusion
- Severity: None
- Status: To be investigated and fixed
- Details: When a user deletes a device, the device is deleted but the GUI shows error message stating "Device was not deleted". NOTE that device being deleted actually means device is marked for deletion and no longer visible for user in GUI. The final device deletion (from the Database) happens on next Dwarfguard service restart however that is only a technicality and the device slot is freed immediately after the device deletion by the user (does not require service restart). The GUI error being shown has no impact on the device deletion process, it is simply a confusing (and wrong) message being shown.
- Workaround: None / Ignore this message
- Alert that is both raised and past is shown only as raised on the dashboard
- Reported: 2026-06-09 for version 2.0.1
- Impact: If an alert is active for some devices, the devices that no longer has this alert triggered but had this alert triggered in past (past alert) are not shown on the Dashboard as an alert is shown as Raised only.
- Severity: Low (once the higher priority stuff (raised alert) is handled, the past alert shows up, correctly listing all affected devices)
- Status: To be investigated