Repository navigation
docs.python.org returning HTTP 503 from multiple Fastly POPs worldwide #158554
Description
I am currently experiencing consistent HTTP 503 responses from https://docs.python.org/.
I have tested the site from three geographically independent locations, using different Fastly points of presence. All three return HTTP 503 responses.
Amsterdam, Netherlands
Direct connection, no VPN:
HTTP/2 503
cache-control: no-cache
content-type: text/html
fastly-restarts: 1
accept-ranges: bytes
date: Thu, 01 Oct 2026 08:30:03 GMT
via: 1.1 varnish
x-served-by: cache-rtm-ehrd2290027-RTM
x-cache: MISS
x-cache-hits: 0
x-timer: S1790843404.501114,VS0,VE318
X-Firefox-Spdy: h2
Toronto, Canada
Firefox through a VPN endpoint in Canada. The VPN location was verified independently using IP leak detection:
HTTP/2 503
cache-control: no-cache
content-type: text/html
fastly-restarts: 1
accept-ranges: bytes
date: Thu, 01 Oct 2026 08:32:23 GMT
via: 1.1 varnish
x-served-by: cache-yyz4557-YYZ
x-cache: MISS
x-cache-hits: 0
x-timer: S1790843544.834536,VS0,VE76
X-Firefox-Spdy: h2
Singapore
Firefox through a VPN endpoint in Singapore:
HTTP/2 503
cache-control: no-cache
content-type: text/html
fastly-restarts: 1
accept-ranges: bytes
date: Thu, 01 Oct 2026 08:33:05 GMT
via: 1.1 varnish
x-served-by: cache-sin-wsat1880058-SIN
x-cache: MISS
x-cache-hits: 0
x-timer: S1790843584.886808,VS0,VE1185
X-Firefox-Spdy: h2
The three requests are being served by different Fastly POPs:
RTM (Amsterdam)
YYZ (Toronto)
SIN (Singapore)
Yet all three consistently return HTTP 503 with:
fastly-restarts: 1
x-cache: MISS
This makes a local network problem or a problem affecting a single Fastly POP unlikely. It appears that the problem may be affecting the docs.python.org service or its backend/origin globally.
Could you please investigate whether there is currently an incident affecting the docs.python.org service or its Fastly backend/origin?
I can provide additional request/response details or test specific documentation URLs if useful.
Activity
hugovk commentedon Oct 1, 2026
Thanks for the report!
This is from an expired certificate, likely caused by an incompatible upgrade of pyOpenSSL, and also affects https://buildbot.python.org/. We're looking into it.
Let's keep this open for now for visibility, as others will come along to report this too :)
alex-wells commentedon Oct 1, 2026
Just to note that https://status.python.org/ is currently reporting the docs infrastructure as operational; depending on what level it's inspecting, I agree that some part of it is operational, despite not being able to access it externally, which might be somewhat confusing to those that haven't seen these threads 😅
StanFromIreland commentedon Oct 1, 2026
Yes, technically everything is working, in that it's at least all up, however since the internal certs disagree the load balancer isn't happy with the docs server.
jayaddison commentedon Oct 1, 2026
I also encountered HTTP 503 errors on
docs.python.orgup until a few minutes ago from different network connections - however it appears to be coming back online now (it returns the expectedPython 3.14.8 DocumentationHTML page, and an HTTP 200 response).StanFromIreland commentedon Oct 1, 2026
This has now been resolved and the docs are back up: https://docs.python.org/3/ 🥳
JulienPalard commentedon Oct 1, 2026
That's not the first time we're having issues with internal certificates. And as far as I remember this is the only reason for downtimes on d.p.o. It should be monitored.
StanFromIreland commentedon Oct 1, 2026
Yes, let's discuss in python/psf-salt#618. We should consider stopping automatic upgrades.
wibek commentedon Oct 1, 2026
I'm happy you guys fixed it so quickly. Great job, thanks.
jayaddison commentedon Oct 1, 2026
Agreed, thank you Python infra/ops maintainers ❤️