翻訳待ち：Show HN: Pingularity – scheduled Ookla/iperf3 speedtest dashboard, outage alerts

AI サービスが一時的に利用できないため、復旧後に翻訳を補完します。

Install→ http://localhost:9000 sudo docker run -d --name pingularity \ --restart unless-stopped \ --network=host --cap-add=NET_RAW \ -v pingularity-data:/var/lib/pingularity \ ghcr.io/pingular/pingularity:latest A self-hosted dashboard for your internet connection. Every outage, speed run, and latency reading, recorded and kept for as long as you want. "Is it down or is it just me?" The call freezes, then Netflix won't load either, and everyone glares at the router. This was running the whole time, so you can just look. Maybe the internet really did drop, and it'll show you when and for how long. Or it never dropped, and now you know it's your Wi-Fi or the laptop, or the website's having a bad day. "Am I getting what I pay for?" You're paying for 500 megs. The last time anyone measured that was probably the day the installer came, between 8 and noon, at 3pm. This runs speed tests through the week, while you're sleeping, or whenever you want it to. If your 500 turns into 80 the same time every night, it'll be sitting right there in the history. "The internet company says it looks fine." They always say that, and the first person on the phone will still make you unplug the router for thirty seconds. Now you've got as much data as you want. Went down four times this month, here are the dates and how long each one lasted. When you escalate or want out of the contract, you've got a list instead of a feeling. Speed tests on a schedule Download, upload, ping, jitter, packet loss, and bufferbloat in both directions - each tracked over time against your plan. Test on Ookla servers or your own iperf3, on a schedule or the moment you hit Run. google 14.3 msgoogle-v6 14.9 mscf 11.9 mscf-v6 12.1 msq9 15.7 msq9-v6 16.8 msdns 18.9 ms Ping and DNS sampled constantly The lowest round-trip to several always-on anchors over IPv4 and IPv6, plus DNS lookup time, sampled every few seconds. Read it from the last five minutes back to a month. Public IP86.75.30.9 (cust-86-75-30-9.paloalto.calnect.net) · 2001:db8:867:5309::9ISPCalNect · Palo Alto, United StatesDNSHooli DNS → 198.51.100.53 (resolver1.hooli.xyz) · Hooli DNS · San Junipero, US iExitmiddle-out.calnect.net (Palo Alto · 3.4 ms · hop 2) → gvn.hooli.net (San Junipero · 4.6 ms · AS64500) · Cloudflare PoP SJC i Know your connection Your public IP and reverse hostname, the ISP and where it sits, the DNS resolver actually answering you (not just the one you configured), and the whole path your traffic takes leaving the ISP - hop by hop, down to the Cloudflare edge. Downtime you can point to Outages are caught by quorum, several landmarks have to agree, so a single lost ping is never a false alarm. Every real one lands on a calendar heatmap, timed to the second, with honest uptime from the last hour to all-time. SpeedtestServerLatencyScheduleAlertsAccessAppearance AutomaticOffOn When degradedOffOn On reconnectOffOn Skip when busyOffOn Run every60 min Degraded above150 ms Configure everything Every part of the monitor is yours to tune - test engines and schedules, alert thresholds, retention, access control - plus nine themes, custom colours, and panels you can rearrange. It all lives in one settings drawer. Alerts A webhook to Discord, Slack, etc. when the internet drops or a threshold is crossed, plus a heartbeat to trigger a notification when the connection is down. Metrics A Prometheus endpoint with an official Grafana dashboard and alert rules ready to import, a plain HTTP API, CSV and full-database export, raw numbers whenever you want them. Private by default One binary, history in a local SQLite file, zero telemetry, login protected, limit access to machine or local network.