mokhtar
released this
2026-10-20 13:57:09 +00:00
| 41 commits to master since this release
Fixed
A blackhole route to an upstream no longer stops failover. On a route like that, connect() fails with an errno the runtime has no name for, and nxdns took it for a fault of its own: the query got SERVFAIL at once, the next upstream was never tried, and nothing was charged to the unreachable one. A connect failure of that kind now counts against that upstream as ConnectFailed (cause Unexpected), and the query fails over.
The Overview charts no longer draw the current bucket as a fall in traffic. Every Overview window ends at the end of the bucket that holds the present moment, so its last bucket is still counting. Queries over time draws the curve into it dashed over a fainter fill, and Client activity draws its column faint inside a dashed outline. Both legends name "Still counting", and the tooltip and the chart's data table give that bucket's time as "
The query detail page names the device. The Client fact showed only the address, while the Activity list showed "iPhone (192.168.50.52)" for the same query. It now uses the list's label. The query log records only the address, so the record's intro says the device name is the one the address has today.
Added
A NixOS module. The flake exports nixosModules.default, which adds services.nxdns with four options: enable, package, configFile and openFirewall. It runs the release's own nxdns.service with a drop-in for file mode, checks configFile with the pinned binary's nxdns import at build time so a configuration that does not parse, validate or reconcile fails the build, and restarts the service when only the configuration changes. The flake's package now also installs lib/systemd/system/nxdns.service. See "Install nxdns with Nix".
nxdns_upstream_local_resource_total counts exchanges that ended on this process's own resources (memory, file descriptors, an unnamed system error) before any upstream answered. Until now, such a SERVFAIL left no record of its cause anywhere.
Changed
The compose file names its own release.deploy/docker/compose.yaml defaults its image to the release it is tagged with instead of :latest, so docker compose up -d with the file from a tag runs that tag's image, and the install and upgrade steps no longer set NXDNS_VERSION. The variable still picks another release, and NXDNS_IMAGE still names a local image. Compose files from v0.0.22 and earlier keep the :latest default.
A DoH upstream that answers with an error status says which. The Diagnostics episode and /api/health read HttpStatus (status 505 HTTP Version Not Supported) instead of HttpStatus (cause HttpStatus). A 505 is a server that requires HTTP/2, which nxdns does not speak.
9a99890 build: bump version to 0.0.23
4ad7ad6 changelog: date the 0.0.23 section
b4ef7b8 website: plan records the approved look, three install tabs and phone screenshots
d9d1c1d docs: verify-a-release transcripts elide the signing key's email address
f351469 nix: a services.nxdns module that runs the packaged unit in file mode against a configuration the pinned binary accepted
aa8c9e2 deploy: the compose file names its own release, and the cut moves it with the version
ea9f16a admin: the overview's last bucket is drawn as still counting; the query detail names the device as the activity list does
8efe504 upstream: an unnamed connect errno fails over, local-resource exchanges are counted, doh faults carry the http status
2697adf website: demo household and the admin screenshots it produced
4ef795a docs: config examples use cloudflare doh; quad9 doh needs http/2 and answers 505
mokhtar released this