mirror / status
Status
all systems normal All three sites are serving and every tree is within its sync interval. Uptime over the last 90 days: 99.97 percent for the service as a whole.
| site | region | uplink | 90 d uptime | p50 first byte | storage | state |
|---|---|---|---|---|---|---|
| bran | eu-north, Brannock campus | 2 x 100 GbE | 99.98 % | 18 ms | 312 TB raw, ZFS, 14 shelves | up |
| fenn | eu-central, colocation | 100 GbE | 99.97 % | 22 ms | 240 TB raw, ZFS, 10 shelves | up |
| kaito | ap-northeast, exchange | 40 GbE | 99.95 % | 31 ms | 128 TB raw, 31 highest-traffic trees | up |
Uptime, last 90 days
bran
90 days agodegraded day 2026-10-02today
fenn
90 days agomaintenance 2026-09-14today
kaito
90 days agoupgrade window 2026-10-12today
Incident history
- 2026-10-12
- kaito uplink upgrade. Scheduled, 02:00 to 03:10 UTC. Traffic served from fenn and bran; no failed requests.
- 2026-10-02
- bran degraded. A disk shelf lost a controller at 11:42 UTC and reads on 9 percent of files slowed from 18 ms to about 400 ms. Resilvering finished at 19:05. No data was lost and geo-DNS moved most clients to fenn within 6 minutes.
- 2026-09-29
- fenn partial outage. A line card on the transit router reset, 14 minutes of packet loss. The vendor replaced it on 2026-10-04.
- 2026-09-14
- fenn maintenance. Planned kernel and firmware updates, 90 minutes, announced seven days ahead.
- 2026-08-27
- kaito peering flap. The exchange moved our port; 22 minutes at reduced capacity.
Service targets
We aim for 99.9 percent monthly availability per site and for every tree to be no more than twice its sync interval behind upstream. These are goals of a volunteer-run service, not a guarantee. If you need a contract, sponsors at the Quarry level and above receive an agreed incident-notification procedure.
Subscribe
Planned maintenance is posted on the news page at least seven days ahead, and unplanned incidents within an hour. Mail abuse@n73.assetlines.net or mirror-admin@n73.assetlines.net to report something not listed here.