Live # Service status. Every component of the platform and every data centre, probed every sixty seconds from five cities. When something breaks we write it here first, and it stays here afterwards: the log goes back to September 2022 and nothing has ever been removed from it. - 99.99% uptime over 90 days, all components - 11 incidents in the last 90 days - 23 components probed every 60 seconds - 1,468 days of history, nothing removed All systems operational 23 components in 6 data centres, probed every 60 seconds from 5 cities. Nothing is open. Last check 21:20 UTC [History](https://gpuserver.io/status/history) ## Components One line per thing that can fail, ninety days of history per line. A bar turns amber for a degradation, red for an outage, blue for an announced maintenance — and it stays that colour. Hover a bar for the day's detail. Platform Shared by every site 99.99% · 90 d Operational Website & configurator gpuserver.io, the catalogue and the order path 99.98% · 90 d Operational 90 days ago 2 day(s) with an outage, 2 day(s) with a degradation or maintenance Today Sign-in & customer area Sessions, orders, machine details, delivery messages 99.99% · 90 d Operational 90 days ago 1 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Payments & invoicing Address issue, confirmation watchers for every coin, term activation 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Provisioning pipeline Image write, driver stack, network configuration, burn-in release 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Support desk Tickets and their notifications; the on-call engineer is separate 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Amsterdam AMS Tier III · 100% wind · our suite since 2021 100% · 90 d Operational Network — transit & peering Cogent, Arelion, Lumen and AMS-IX, NL-ix 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 4 day(s) with a degradation or maintenance Today Power & cooling Tier III, 100% wind, dual-fed racks 100% · 90 d Operational 90 days ago 90 days without an incident Today Out-of-band management IPMI, KVM console, remote power, reinstalls 100% · 90 d Operational 90 days ago 90 days without an incident Today Stockholm STO Tier III · 100% hydro · our suite since 2022 100% · 90 d Operational Network — transit & peering Arelion, GlobalConnect and Netnod, STHIX 100% · 90 d Operational 90 days ago 90 days without an incident Today Power & cooling Tier III, 100% hydro, dual-fed racks 100% · 90 d Operational 90 days ago 90 days without an incident Today Out-of-band management IPMI, KVM console, remote power, reinstalls 100% · 90 d Operational 90 days ago 90 days without an incident Today Zürich ZRH Tier IV · 96% carbon-free · our suite since 2023 100% · 90 d Operational Network — transit & peering Init7, Cogent and SwissIX, CIXP 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Power & cooling Tier IV, 96% carbon-free, dual-fed racks 100% · 90 d Operational 90 days ago 90 days without an incident Today Out-of-band management IPMI, KVM console, remote power, reinstalls 100% · 90 d Operational 90 days ago 90 days without an incident Today Reykjavík REY Tier III · 100% geothermal + hydro · our suite since 2024 99.99% · 90 d Operational Network — transit & peering Farice, Síminn and RIX 99.99% · 90 d Operational 90 days ago 1 day(s) with an outage, 2 day(s) with a degradation or maintenance Today Power & cooling Tier III, 100% geothermal + hydro, dual-fed racks 100% · 90 d Operational 90 days ago 90 days without an incident Today Out-of-band management IPMI, KVM console, remote power, reinstalls 99.99% · 90 d Operational 90 days ago 1 day(s) with an outage, 0 day(s) with a degradation or maintenance Today Montréal YUL Tier III · 99% hydro · our suite since 2023 100% · 90 d Operational Network — transit & peering Cogent, Zayo and QIX 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 3 day(s) with a degradation or maintenance Today Power & cooling Tier III, 99% hydro, dual-fed racks 100% · 90 d Operational 90 days ago 0 day(s) with an outage, 1 day(s) with a degradation or maintenance Today Out-of-band management IPMI, KVM console, remote power, reinstalls 100% · 90 d Operational 90 days ago 90 days without an incident Today Singapore SIN Tier III · Grid + REC offset · our suite since 2025 99.99% · 90 d Operational Network — transit & peering Arelion, Singtel and SGIX, Equinix Singapore 99.99% · 90 d Operational 90 days ago 1 day(s) with an outage, 0 day(s) with a degradation or maintenance Today Power & cooling Tier III, Grid + REC offset, dual-fed racks 100% · 90 d Operational 90 days ago 90 days without an incident Today Out-of-band management IPMI, KVM console, remote power, reinstalls 100% · 90 d Operational 90 days ago 90 days without an incident Today - Operational - Degraded performance - Partial outage - Major outage - Maintenance - Not yet in service ## Response time Time to first byte of gpuserver.io over HTTPS, measured every minute from five cities and plotted one point per five minutes. A platform incident shows here as a bump; a red band is a period when the site did not answer at all. Median over the period: 186 ms · 95th percentile: 358 ms · 289 points per city ## Past incidents The last fourteen days, day by day, including the days when nothing happened. Older events are in the history, month by month, back to September 2022. ### 11 September 2026 today No incidents reported. ### 10 September 2026 No incidents reported. ### 9 September 2026 No incidents reported. ### 8 September 2026 No incidents reported. ### 7 September 2026 No incidents reported. ### 6 September 2026 No incidents reported. ### 5 September 2026 No incidents reported. ### 4 September 2026 No incidents reported. ### 3 September 2026 No incidents reported. ### 2 September 2026 No incidents reported. ### 1 September 2026 Maintenance Completed ### [Payment watcher node upgrade (Ethereum)](https://gpuserver.io/status/incidents/67hh86xc66) Payments & invoicing 02:00–03:50 UTC · 1 h 50 min **Completed** — The Ethereum node is on Geth 1.17.0 and at the tip. Every payment made during the window has been credited. ### 31 August 2026 No incidents reported. ### 30 August 2026 No incidents reported. ### 29 August 2026 No incidents reported. [Full incident history](https://gpuserver.io/status/history) ## Scheduled maintenance Announced here, in the feeds and in the calendar at least five days ahead, at night in the site's own time zone, with the exact impact spelled out. An announced interruption still counts against the SLA of the machines it touches. Nothing scheduled in the next 28 days. When a window is announced it appears here first. Subscribe to the calendar to have it land in your own. Completed in the last 45 days - [1 Sept 2026 Payment watcher node upgrade (Ethereum) 1 h 50 min](https://gpuserver.io/status/incidents/67hh86xc66) - [26 Aug 2026 Website runtime upgrade (PHP 8.5) 25 minutes](https://gpuserver.io/status/incidents/gph8pxz8ja) - [29 Jul 2026 Core router software upgrade — YUL 1 h 53 min](https://gpuserver.io/status/incidents/4zqur5sbzz) - [29 Jul 2026 Top-of-rack switch firmware, rows D–E — AMS 1 h 59 min](https://gpuserver.io/status/incidents/t2ms25xr82) ## How this page is measured A status page is only worth what its probes are worth. Here is what runs, from where, and what it takes for a state to change. ### Probes, every 60 seconds HTTPS, TCP and BGP checks run once a minute from Frankfurt, London, New York, Singapore and São Paulo — outside our own network, on purpose. A state changes only when at least three of the five vantage points agree, which filters a single bad probe without hiding a real fault for more than two minutes. ### What counts as downtime Partial and major outages count against uptime, minute by minute. Degraded performance and announced maintenance do not — they are drawn on the bars but never removed from them. The service level agreement is settled on exactly this data, per machine, without a claim. ### Who writes the updates The engineer working the incident, in English, at the time — never rewritten afterwards. Every major or critical incident gets a post-incident review within five days, with the cause and what changed. The log goes back to 5 September 2022 and nothing has been removed from it. ### Take it with you Every event on this page is published at the same moment to the feeds below. Point your monitoring, your chat tool or your calendar at them; no account is needed. - [**RSS** /status/rss](https://gpuserver.io/status/rss) - [**Atom** /status/atom](https://gpuserver.io/status/atom) - [**JSON API** /status/api/v2/summary.json](https://gpuserver.io/status/api/v2/summary.json) - [**Maintenance calendar** /status/maintenance.ics](https://gpuserver.io/status/maintenance.ics) - [**Status badge** /status/badge.svg](https://gpuserver.io/status/badge.svg) ## What an outage is worth to you Not goodwill, and not an account credit — there is no balance here to credit. Downtime is repaid in time on the machine you rented, applied without a claim. 99.9% Per machine, per calendar month Measured on network reachability and power, from outside the suite. Roughly 43 minutes a month. ×5 The time back, five times over An hour lost extends your term by five hours, to a cap of one full month per calendar month. 0 Forms to fill in We apply it from our own measurements. You are told after the fact, not asked to prove it. The exclusions — scheduled work you were told about, anything caused inside your own operating system, and a force majeure clause — are written out in full on the [service level agreement](https://gpuserver.io/legal/sla). They are worth two minutes before you need them. ## Nothing above is a reason to trust us. It is a reason to check. The probes run whether anyone is looking or not, and this page shows what they found — including the days they found nothing. [Browse the catalogue](https://gpuserver.io/#catalog) [Read the guides](https://gpuserver.io/guides) ## Related pages - [Incident history Every incident and maintenance window on gpuserver.io since monitoring began in September 2022, month by month, with its timeline and post-incident review.](https://gpuserver.io/status/history) - [Service level agreement The 99.9% commitment: what counts as downtime, how it is measured from outside, five minutes of term back per minute lost, and the exclusions in full.](https://gpuserver.io/legal/sla) - [Network Unmetered ports up to 25 Gbit/s, two carriers and an IX per site, always-on DDoS filtering, routed IPv6 — and the four things we do not offer, stated up front.](https://gpuserver.io/network) - [Hardware policy The twelve NVIDIA GPUs we operate, the 72-hour burn-in before a node is sold, the four-hour replacement target, and how disks are erased between tenants.](https://gpuserver.io/hardware) --- Source: https://gpuserver.io/status/. This file is generated from the same data as the website; if a figure here differs from a page, the page is authoritative and this file is stale — the canonical source is https://gpuserver.io/.