Skip to content

Outage tracker / Fly.io

Is Fly.io down?

I'm having a problem

Operationalchecked moments ago

We reached Fly.io from all 10 regions we check.

Reachability verified

No outage recorded in the 14 days we have measured Fly.io (since Aug 22, 2026).

our independent probeslive · checked Sep 5, 2026, 2:07 AM UTC
US-East499 msOperational
US-West267 msOperational
Europe278 msOperational
Asia-Pacific238 msOperational
US-Central460 msOperational
Canada287 msOperational
UK93 msOperational
Southeast Asia406 msOperational
Australia874 msOperational
South America677 msOperational
their status pagetheir claim, not our reading

Fly.io reports: All systems operational (“All Systems Operational”)

from Fly.io's own status page, checked Sep 5, 2026, 1:46 AM UTC
Our probes and Fly.io's own status page agree: both show Fly.io as reachable right now.

Fly.io's own component breakdown, as they publish it

  • AMS - Amsterdam, NetherlandsOperational
  • Customer ApplicationsOperational
  • Fly Machine .internal DNSOperational
  • Management Plane - ORDOperational
  • Upstash for RedisOperational
  • Usage Metrics APIOperational
  • ARN - Stockholm, SwedenOperational
  • DashboardOperational
  • Fly Machine External DNSOperational
  • Management Plane - IADOperational
  • Stripe API ConnectionOperational
  • Machines APIOperational
  • Management Plane - FRAOperational
  • *.flyio.net NameserversOperational
  • Management Plane - GRUOperational
  • BOM - Mumbai, IndiaOperational
  • flydns.netOperational
  • Management Plane - LAXOperational
  • Persistent Storage (Volumes)Operational
  • CDG - Paris, FranceOperational
  • DeploymentsOperational
  • Management Plane - SYDOperational
  • Management Plane - AMSOperational
  • Remote BuildsOperational
  • DFW - Dallas, Texas (US)Operational
  • LogsOperational
  • Management Plane - LHROperational
  • EWR - Secaucus, NJ (US)Operational
  • Management Plane - NRTOperational
  • MetricsOperational
  • Management Plane - SINOperational
  • SSL/TLS Certificate ProvisioningOperational
  • Anycast EdgeOperational
  • FRA - Frankfurt, GermanyOperational
  • Management Plane - SJCOperational
  • Fly Machine Image Registry 1Operational
  • Management Plane - YYZOperational
  • Fly Machine Image Registry 2Operational
  • GRU - Sao Paulo, BrazilOperational
  • IAD - Ashburn, Virginia (US)Operational
  • CorrosionOperational
  • JNB - Johannesburg, South AfricaOperational
  • LAX - Los Angeles, California (US)Operational
  • LHR - London, United KingdomOperational
  • Phoenix.newOperational
  • Support PortalOperational
  • SpritesOperational
  • NRT - Tokyo, JapanOperational
  • ORD - Chicago, Illinois (US)Operational
  • SIN - SingaporeOperational
  • SJC - San Jose, California (US)Operational
  • SYD - Sydney, AustraliaOperational
  • YYZ - Toronto, CanadaOperational

These are Fly.io's components, named and scored by Fly.io. We do not map them onto the functions we check above, because we have no way to know which of their components our checks touch. Two lists, side by side, neither one rewriting the other.

This is Fly.io's own claim about their service, shown separately from what we measured. We never change our own numbers based on what a vendor reports, and this reading never changes theirs. Read both.

Fly.io response time, by region

US-East
537msp95 · latest
2s1s0ms
US-East response time (p95) over the last 30 daysDaily 95th-percentile response time for US-East, scaled to a maximum of 2s. Days with no samples are not plotted, not shown as zero.2026-08-22 UTC: p95 440ms2026-08-23 UTC: p95 220ms2026-08-24 UTC: p95 529ms2026-08-25 UTC: p95 676ms2026-08-26 UTC: p95 495ms2026-08-27 UTC: p95 424ms2026-08-28 UTC: p95 468ms2026-08-29 UTC: p95 1.3s2026-08-30 UTC: p95 1.3s2026-08-31 UTC: p95 800ms2026-09-01 UTC: p95 586ms2026-09-02 UTC: p95 524ms2026-09-03 UTC: p95 995ms2026-09-04 UTC: p95 1.0s2026-09-05 UTC: p95 537ms
Aug 7Aug 21Sep 5
US-West
388msp95 · latest
1s500ms0ms
US-West response time (p95) over the last 30 daysDaily 95th-percentile response time for US-West, scaled to a maximum of 1s. Days with no samples are not plotted, not shown as zero.2026-08-22 UTC: p95 366ms2026-08-23 UTC: p95 242ms2026-08-24 UTC: p95 395ms2026-08-25 UTC: p95 291ms2026-08-26 UTC: p95 256ms2026-08-27 UTC: p95 686ms2026-08-28 UTC: p95 519ms2026-08-29 UTC: p95 520ms2026-08-30 UTC: p95 715ms2026-08-31 UTC: p95 366ms2026-09-01 UTC: p95 355ms2026-09-02 UTC: p95 493ms2026-09-03 UTC: p95 426ms2026-09-04 UTC: p95 462ms2026-09-05 UTC: p95 388ms
Aug 7Aug 21Sep 5
Europe
448msp95 · latest
1s500ms0ms
Europe response time (p95) over the last 30 daysDaily 95th-percentile response time for Europe, scaled to a maximum of 1s. Days with no samples are not plotted, not shown as zero.2026-08-22 UTC: p95 443ms2026-08-23 UTC: p95 326ms2026-08-24 UTC: p95 528ms2026-08-25 UTC: p95 564ms2026-08-26 UTC: p95 206ms2026-08-27 UTC: p95 338ms2026-08-28 UTC: p95 392ms2026-08-29 UTC: p95 361ms2026-08-30 UTC: p95 275ms2026-08-31 UTC: p95 394ms2026-09-01 UTC: p95 530ms2026-09-02 UTC: p95 328ms2026-09-03 UTC: p95 252ms2026-09-04 UTC: p95 456ms2026-09-05 UTC: p95 448ms
Aug 7Aug 21Sep 5
Asia-Pacific
681msp95 · latest
2s1s0ms
Asia-Pacific response time (p95) over the last 30 daysDaily 95th-percentile response time for Asia-Pacific, scaled to a maximum of 2s. Days with no samples are not plotted, not shown as zero.2026-08-22 UTC: p95 456ms2026-08-23 UTC: p95 777ms2026-08-24 UTC: p95 777ms2026-08-25 UTC: p95 847ms2026-08-26 UTC: p95 719ms2026-08-27 UTC: p95 865ms2026-08-28 UTC: p95 745ms2026-08-29 UTC: p95 751ms2026-08-30 UTC: p95 564ms2026-08-31 UTC: p95 636ms2026-09-01 UTC: p95 684ms2026-09-02 UTC: p95 666ms2026-09-03 UTC: p95 1.2s2026-09-04 UTC: p95 1.2s2026-09-05 UTC: p95 681ms
Aug 7Aug 21Sep 5
US-Central
544msp95 · latest
1s500ms0ms
US-Central response time (p95) over the last 30 daysDaily 95th-percentile response time for US-Central, scaled to a maximum of 1s. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 540ms2026-09-04 UTC: p95 657ms2026-09-05 UTC: p95 544ms
Aug 7Aug 21Sep 5
Canada
359msp95 · latest
500ms250ms0ms
Canada response time (p95) over the last 30 daysDaily 95th-percentile response time for Canada, scaled to a maximum of 500ms. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 439ms2026-09-04 UTC: p95 357ms2026-09-05 UTC: p95 359ms
Aug 7Aug 21Sep 5
UK
175msp95 · latest
500ms250ms0ms
UK response time (p95) over the last 30 daysDaily 95th-percentile response time for UK, scaled to a maximum of 500ms. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 336ms2026-09-04 UTC: p95 218ms2026-09-05 UTC: p95 175ms
Aug 7Aug 21Sep 5
Southeast Asia
745msp95 · latest
1s500ms0ms
Southeast Asia response time (p95) over the last 30 daysDaily 95th-percentile response time for Southeast Asia, scaled to a maximum of 1s. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 812ms2026-09-04 UTC: p95 923ms2026-09-05 UTC: p95 745ms
Aug 7Aug 21Sep 5
Australia
1.3sp95 · latest
2s1s0ms
Australia response time (p95) over the last 30 daysDaily 95th-percentile response time for Australia, scaled to a maximum of 2s. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 1.4s2026-09-04 UTC: p95 1.4s2026-09-05 UTC: p95 1.3s
Aug 7Aug 21Sep 5
South America
660msp95 · latest
1s500ms0ms
South America response time (p95) over the last 30 daysDaily 95th-percentile response time for South America, scaled to a maximum of 1s. Days with no samples are not plotted, not shown as zero.2026-09-03 UTC: p95 865ms2026-09-04 UTC: p95 823ms2026-09-05 UTC: p95 660ms
Aug 7Aug 21Sep 5

Is it Fly.io, or is it you?

Watching an outage above does not tell you whether your own connection or site is also having a problem. Enter your own URL for one live check, right now, no signup. This checks the site you enter, not Fly.io -- it cannot diagnose Fly.io's own systems.

What are you seeing? (optional)

No account needed, and we never store your IP address. Reports are counted separately from our own measurements and are always shown as unverified.

What we measured, and what people reported

  • North America
    OperationalUS-East, US-West, US-Central, Canada
    • US-East
      Operationallatency: 499msstatus code: 200checked: Sep 5, 2026, 2:07 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur US-East probes reach Fly.io normally. We do not have enough reports from US-East to say anything about what people there are seeing.
    • US-West
      Operationallatency: 267msstatus code: 200checked: Sep 5, 2026, 2:03 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur US-West probes reach Fly.io normally. We do not have enough reports from US-West to say anything about what people there are seeing.
    • US-Central
      Operationallatency: 460msstatus code: 200checked: Sep 5, 2026, 2:03 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur US-Central probes reach Fly.io normally. We do not have enough reports from US-Central to say anything about what people there are seeing.
    • Canada
      Operationallatency: 287msstatus code: 200checked: Sep 5, 2026, 1:53 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur Canada probes reach Fly.io normally. We do not have enough reports from Canada to say anything about what people there are seeing.
  • South America
    Operationallatency: 677msstatus code: 200checked: Sep 5, 2026, 1:56 AM UTC
    0 people in the last 5 minutesNot enough reports to sayOur South America probes reach Fly.io normally. We do not have enough reports from South America to say anything about what people there are seeing.
  • Europe
    OperationalEurope, UK
    • Europe
      Operationallatency: 278msstatus code: 200checked: Sep 5, 2026, 2:06 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur Europe probes reach Fly.io normally. We do not have enough reports from Europe to say anything about what people there are seeing.
    • UK
      Operationallatency: 93msstatus code: 200checked: Sep 5, 2026, 1:58 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur UK probes reach Fly.io normally. We do not have enough reports from UK to say anything about what people there are seeing.
  • Asia Pacific
    OperationalAsia-Pacific, Southeast Asia, Australia
    • Asia-Pacific
      Operationallatency: 238msstatus code: 200checked: Sep 5, 2026, 2:02 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur Asia-Pacific probes reach Fly.io normally. We do not have enough reports from Asia-Pacific to say anything about what people there are seeing.
    • Southeast Asia
      Operationallatency: 406msstatus code: 200checked: Sep 5, 2026, 1:47 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur Southeast Asia probes reach Fly.io normally. We do not have enough reports from Southeast Asia to say anything about what people there are seeing.
    • Australia
      Operationallatency: 874msstatus code: 200checked: Sep 5, 2026, 2:07 AM UTC
      0 people in the last 5 minutesNot enough reports to sayOur Australia probes reach Fly.io normally. We do not have enough reports from Australia to say anything about what people there are seeing.

Somewhere else or not sure: 0 people in the last 5 minutes

These are unverified reports from people using Fly.io, counted separately from what our own probes measure. Anyone can file one, no account is needed, and the region on a report is chosen by the person filing it. We never open, extend or close an outage in our record based on report volume: only our own measurements do that. We have not received any reports for Fly.io yet, which is not evidence that nothing is wrong.

Recent outages

Sprites API Partial Outage, 1h 38mdeclared on their status page2026-09-02
Upstream network issues, 1h 39mdeclared on their status page2026-09-02
API background job queue failure, 23mdeclared on their status page2026-09-02
HTTP/2 traffic disruptions, 1mdeclared on their status page2026-08-31
Packet loss in ORD, 1h 44mdeclared on their status page2026-08-31
Sprite deletion jobs failing, 1mdeclared on their status page2026-08-30
Networking Issues in GRU, 40mdeclared on their status page2026-08-28
Increased packet loss, 2h 14mdeclared on their status page2026-08-28

measured entries are our own probes' readings; declared entries are the vendor's. How this check works

We have been measuring Fly.io since August 22, 2026. We show nothing from before that date, and we never add an outage we did not observe ourselves, from any source.

Need this in writing? Generate an SLA evidence report for Fly.io covering any window you pick, with our measurements and Fly.io's own claim at the time, side by side.

Full availability history for Fly.io: 90-day availability by region, month-by-month uptime, and every incident on record.

Monitor your own site the same way

This page shows what we see when we check this service, using the same ten regions, for a site you run. RealUptime Monitor watches your own servers from inside your network, and it starts free.

Start monitoring your own servers

Tell me when this goes down

Get told when we measure this

We will tell you when our probes stop reaching Fly.io, and again when they can reach it again. That is our own measurement from ten regions, not Fly.io's status page. No account, no card.

Delivery channel

Double opt-in: nothing is sent until you click the link in the confirmation email. One email when a measured outage opens, one when it closes, and a one-click unsubscribe in every one. We do not send anything else to this address.

Put this badge in your README

Fly.io uptime measured by RealUptimeLive, from this page's own readings.
[![Fly.io uptime measured by RealUptime](https://realuptime.io/outages/fly-io/badge.svg?days=30)](https://realuptime.io/outages/fly-io)
<a href="https://realuptime.io/outages/fly-io"><img src="https://realuptime.io/outages/fly-io/badge.svg?days=30" alt="Fly.io uptime measured by RealUptime"></a>

The badge shows the share of our probes that succeeded over the last 30 days, and the window is printed on it so the number always says what period it covers. If our readings go stale it says stale, and if we have no readings it says so: it will not show green on data we cannot vouch for. Free, no account, and it links back here. See how the badge works for every window and the honesty rules behind it.

What Fly.io's SLA implies, from our measurement

At or above your target

We measured none of downtime for Fly.io over the 338h 8m since we started watching it on 2026-08-22, which is 100% uptime by our reading. A 99.9% target allows 20 minutes over that period.

By our measurement there is no shortfall to claim for this period.

This is computed from our own probe measurements, not from Fly.io's SLA instrumentation. Their contract is normally measured against their own systems, with their own exclusions and a claim window, so their figure and ours will differ. Use this to start the conversation and ask for their number. It is not legal advice and it is not a determination of what you are owed.

What we check

primary
https://fly.io/
secondary
https://status.flyio.net/api/v2/status.json
RealUptime is not affiliated with Fly.io; the name identifies what we measured, and nothing here speaks for Fly.io's own team. We send a single web request to Fly.io from 10 regions on our own schedule and time the response. We do not see inside Fly.io's systems, do not use an account or real transaction, and are not the same signal Fly.io's own team has access to. Treat this as one external data point, not a verdict. What we check for Fly.io proves reachability only: the endpoints above answer, but that does not confirm the underlying product or account features work. Read the full methodology.