BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//gpuserver.io//status//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:gpuserver.io maintenance
X-WR-TIMEZONE:UTC
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-PUBLISHED-TTL:PT1H
BEGIN:VEVENT
UID:67hh86xc66@gpuserver.io
DTSTAMP:20260826T203800Z
DTSTART:20260901T020000Z
DTEND:20260901T035000Z
SUMMARY:[gpuserver.io] Payment watcher node upgrade (Ethereum)
DESCRIPTION:The Ethereum node behind our payment watcher is upgraded to Ge
 th 1.17.0. Detection of Ethereum payments pauses while the node restarts 
 and catches up\; every payment made during the window is detected and cre
 dited afterwards\, with the term starting from the actual confirmation ti
 me. The other coins are unaffected.\n\nAffected: Payments & invoicing\nSt
 atus: Completed\nhttps://gpuserver.io/status/incidents/67hh86xc66
URL:https://gpuserver.io/status/incidents/67hh86xc66
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:gph8pxz8ja@gpuserver.io
DTSTAMP:20260818T190000Z
DTSTART:20260826T040000Z
DTEND:20260826T042500Z
SUMMARY:[gpuserver.io] Website runtime upgrade (PHP 8.5)
DESCRIPTION:The website and customer area move to PHP 8.5. Expect a few se
 conds of errors while the application restarts\, twice. Payments in progr
 ess are unaffected: the watchers do not run on this host. If you are on t
 he payment page\, it will simply reload.\n\nAffected: Website & configura
 tor\, Sign-in & customer area\nStatus: Completed\nhttps://gpuserver.io/st
 atus/incidents/gph8pxz8ja
URL:https://gpuserver.io/status/incidents/gph8pxz8ja
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:4zqur5sbzz@gpuserver.io
DTSTAMP:20260717T062400Z
DTSTART:20260729T084500Z
DTEND:20260729T103800Z
SUMMARY:[gpuserver.io] Core router software upgrade — YUL
DESCRIPTION:We will upgrade the operating system on the two core routers i
 n YUL\, one after the other. Each router is drained before it reloads\, s
 o traffic stays on the other one. Expect up to 60 seconds of reduced redu
 ndancy per router while its BGP sessions reconverge\; no machine should l
 ose connectivity. Out-of-band access is unaffected.\n\nAffected: Network 
 — transit & peering — YUL\nStatus: Completed\nhttps://gpuserver.io/st
 atus/incidents/4zqur5sbzz
URL:https://gpuserver.io/status/incidents/4zqur5sbzz
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:t2ms25xr82@gpuserver.io
DTSTAMP:20260720T232600Z
DTSTART:20260729T014500Z
DTEND:20260729T034400Z
SUMMARY:[gpuserver.io] Top-of-rack switch firmware\, rows D–E — AMS
DESCRIPTION:The top-of-rack switches in rows D to E of AMS receive a firmw
 are update. Each switch reloads once\, which interrupts the network for t
 he machines in its rack for about 90 seconds. Racks are done one at a tim
 e. Running workloads are not affected\, only reachability during the relo
 ad. The interruption is counted against the SLA of the affected machines 
 even though it is announced.\n\nAffected: Network — transit & peering 
 — AMS\nStatus: Completed\nhttps://gpuserver.io/status/incidents/t2ms25x
 r82
URL:https://gpuserver.io/status/incidents/t2ms25xr82
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:szwphghzrr@gpuserver.io
DTSTAMP:20260710T195500Z
DTSTART:20260721T020000Z
DTEND:20260721T041400Z
SUMMARY:[gpuserver.io] Optical transport maintenance on the AMS-IX port 
 — AMS
DESCRIPTION:The exchange operator is carrying out maintenance on the optic
 al transport behind our AMS-IX port in AMS. Peering sessions over AMS-IX 
 will be down for the window\; traffic to those peers goes via transit ins
 tead\, a few milliseconds further. No machine is affected.\n\nAffected: N
 etwork — transit & peering — AMS\nStatus: Completed\nhttps://gpuserve
 r.io/status/incidents/szwphghzrr
URL:https://gpuserver.io/status/incidents/szwphghzrr
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:ckdfaedfkr@gpuserver.io
DTSTAMP:20260704T232700Z
DTSTART:20260715T080000Z
DTEND:20260715T100700Z
SUMMARY:[gpuserver.io] Top-of-rack switch firmware\, rows A–B — YUL
DESCRIPTION:The top-of-rack switches in rows A to B of YUL receive a firmw
 are update. Each switch reloads once\, which interrupts the network for t
 he machines in its rack for about 90 seconds. Racks are done one at a tim
 e. Running workloads are not affected\, only reachability during the relo
 ad. The interruption is counted against the SLA of the affected machines 
 even though it is announced.\n\nAffected: Network — transit & peering 
 — YUL\nStatus: Completed\nhttps://gpuserver.io/status/incidents/ckdfaed
 fkr
URL:https://gpuserver.io/status/incidents/ckdfaedfkr
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:33ud45wpkc@gpuserver.io
DTSTAMP:20260610T051500Z
DTSTART:20260624T130000Z
DTEND:20260624T140400Z
SUMMARY:[gpuserver.io] Facility generator and transfer-switch test — YUL
DESCRIPTION:The facility runs its periodic test of the generators and auto
 matic transfer switches in YUL. The site is transferred to generator and 
 back under load. The UPS covers the transfers\; machines do not notice. W
 e post it because the site is on a single source for a few minutes during
  each transfer.\n\nAffected: Power & cooling — YUL\nStatus: Completed\n
 https://gpuserver.io/status/incidents/33ud45wpkc
URL:https://gpuserver.io/status/incidents/33ud45wpkc
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:dsde29y3p7@gpuserver.io
DTSTAMP:20260609T163900Z
DTSTART:20260624T013000Z
DTEND:20260624T034500Z
SUMMARY:[gpuserver.io] Core router software upgrade — ZRH
DESCRIPTION:We will upgrade the operating system on the two core routers i
 n ZRH\, one after the other. Each router is drained before it reloads\, s
 o traffic stays on the other one. Expect up to 60 seconds of reduced redu
 ndancy per router while its BGP sessions reconverge\; no machine should l
 ose connectivity. Out-of-band access is unaffected.\n\nAffected: Network 
 — transit & peering — ZRH\nStatus: Completed\nhttps://gpuserver.io/st
 atus/incidents/dsde29y3p7
URL:https://gpuserver.io/status/incidents/dsde29y3p7
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
BEGIN:VEVENT
UID:gaxmfa365x@gpuserver.io
DTSTAMP:20260609T180400Z
DTSTART:20260618T034500Z
DTEND:20260618T044100Z
SUMMARY:[gpuserver.io] Provisioning image refresh (Ubuntu 26.04 LTS\, driv
 er 590.44.02)
DESCRIPTION:We publish refreshed operating-system images with the current 
 NVIDIA driver (590.44.02). New provisioning is paused for the window whil
 e the images are validated on one node of each GPU model. Orders that arr
 ive during the window are delivered right after it closes. Machines alrea
 dy delivered are not touched.\n\nAffected: Provisioning pipeline\nStatus:
  Completed\nhttps://gpuserver.io/status/incidents/gaxmfa365x
URL:https://gpuserver.io/status/incidents/gaxmfa365x
STATUS:CONFIRMED
CATEGORIES:MAINTENANCE
END:VEVENT
END:VCALENDAR