Forums

Im running code on webapp and also a schedule task. Nothing is running, Unable to connect to PA. Website is also down.

When you cant access even their web pages, that usually means somebody is attacking the server or even the whole cluster of servers.

Or, your internet connection or provider is futzed.

We can confirm that we encountered the same issue. It was an issue with their loadbalancer, but the apps are still running “slow.”

bro this is getting ridiculous, its like once a month for a few days everything goes down. im still down

I'm also having some issues with consoles running extra slow even though were not "in the tar pit".

My flask web app has been down all day

My file navigation has been so slow. Bash working find, refresh web fine. Just moving between folders and uploading files is very slow.

Web app down / unresponsive — mtsdu.pythonanywhere.com I'm also experiencing this issue. My web app (mtsdu.pythonanywhere.com) has been unreachable since around July 19, 2026, 05:00 UTC+7. Requests just hang indefinitely with no response (no timeout error, no HTTP error code — just infinite loading). Diagnostics from my side:

CPU usage: only 8% used, well within quota File storage: only 6% used, well within quota Error log: no relevant errors around the time of the outage Server log: uWSGI shows the app started successfully and was serving requests normally curl from within my own Bash console to my own app also times out completely

Everything on my account looks healthy, so this seems to match the load balancer issue others have reported in this thread. Just adding my case for visibility. Hope this gets resolved soon.

My webpage will not load and I'm getting a 504 load balancer error

anyone else still down? its been more than 3 hours! i cant access anything

OK so its not just me and/or my application? Everything on my web app feels slower than before. So it's just wait untill this is fixed? Of is there something I can do?

@DNZB, since the load balancer issue last Friday, we’ve also noticed a decline in the performance of our web applications. Although the situation has improved since Friday, it still feels sluggish and slow. For instance, endpoints that used to take 25-100 milliseconds now take 1-2 seconds to respond.

P.S. I’d be more than willing to pay a premium for PA if it meant we could avoid this issue altogether!

@ xDev thanks for your response. But still wondering if there is anything I can do in my application to make things better..? Or is there any status updates from PA about this?

I also have this problem. Cannot even reload my application as it fails..

We had an issue with one of our storage servers that was causing issues across the cluster. We have repaired it and everything should be working again. We will continue to monitor and respond to issues, while we try to fix it

Yeah, it does not seem to have solved the issue. I'm still having major problems. Nothing loads, no requests comes through and I cannot reload my web app successfully.

@glenn, is there anyway we can tell if you guys have fixed all the issues? The response times in my webapplication are going from speed of light to snailspeed and are inconsistent as hell.

@DNZB Email us with timestapmed examples so we could compare with what we see on our side.

@fjl It was just down again a few hours ago till now, so what's the problem cause ?

The PythonAnywhere website was offline for over four hours. What was the problem? The same issue occurred on the 17th.

Absolute joke. FOUR hours where we can't even access the pythonanywhere website to voice our complaints!?

I'm still having problems with one webapp, I have other webapp working normally again after many hours but it seems that it's not totally solved

I'm curious how deep these issues are. For the past maybe 2 weeks there have been intermittent outages, gawdawful response times and then just wonky issues where everything SEEMS to be working, but web pages don't render the data that all the debug logs and troubleshooting (curls etc) show that they should be rendering. Everything was down again earlier today (maybe 8 hours ago) and now it's back to slow and data not rendering on the pages.

Yeah, my web app is down too. Again.

At a bare minimum, there needs to be some way to get notifications about this. I don't think the twitter/X feed is active anymore (which I support).

Hilariously, I can't even "Subscribe to this topic" because of a server error.

Hilariously, I can't even "Subscribe to this topic" because of a server error.

Hilariously, I can't even "Subscribe to this topic" because of a server error.

Hilariously, I can't even "Subscribe to this topic" because of a server error.

"We stand things up and we shut them down" might be the right thought process here.... sigh

Relax, everyone; I’ve been using their service for years, and it’s always gone smoothly. There must be some issue going on these days. But everything is back to normal here now.

Looks like hopefully we are done with the issues.

Congrats to the devs on the latest outage. You've lost a customer

Yea, after the three hour downtime last week I started planning to move to Render. I just moved everything over.

I think PA is still a great service for many things, just not running my SaaS business on.

I really like PythonAnywhere; I've been using it for two years. It actually went down again for an hour today; I hope they can fix this—and the performance issues—once and for all.

down once more used for over two years this last month has not been to good. we down for the second time today.

wondering if this is a me issue or others having issues now?

my websites are still very slow even right now, sometimes not loading, including the pythonanywhere website itself.

seems to be an everyday thing now.

wondering what's changed this last month to have such issues? we are left in dark many times about it not cool!

up down up down. gotta stay up

Agree still down for me slow website cant use my app.

No email when down time happens or why!

No compensation for down time.

Down, all the time now! Fix, please!

@Pythonanywhere Staff: Could you please explain what's so hard about letting us know about outages? I think everyone would understand. But I—and many others—don't understand why you can't seem to manage to inform customers about them. I mean, I pay $100 every month for this service. I don’t expect it to be available 24 hours a day, every day. But I do think it’s reasonable to expect to at least receive a very brief email in advance so that we can implement workarounds. I deal with many other cloud providers in my line of work. When it comes to reliability, they're no better. But every single one of them manages to notify customers immediately about outages.

As it stands now, your service is definitely not suitable for commercial use, which is why I'm unfortunately looking for another provider at the moment. That's a real shame, because your services are actually very good and everything works great. Is there any way, for example, to run an independent VIM that simply checks every second to see if it can access your homepage? If not, just send the data to some email API or X-API, and that's it. I could try building that myself, but I think you'd need admin privileges to scrape your homepage.

Their service has always been 24/7. However, it turns out that even PA's own website goes down for hours, right along with the customers' sites. They need to fix this problem as soon as possible, because it didn't use to be like this.

still experiencing extreme slowness and responsiveness on the main website

have you solved the issue? pa private area is working but my applications are not :/ do you still have problems? thanks!

Dealing with problems from pythonanywhere has become another job of it's own.

백날천날 서버 날아가는 건 그렇다치고, 왜 내 도메인에 남의 사이트가 표시되는 겁니까? I used pythonanywhere almost 6 years, but service of this month really awful.

Still seeing this today (27 July) — looks like the same storage-cluster issue

Chiming in here as I think I'm hitting the same problem described in this thread. My web app has been effectively offline for several hours this morning (UTC) with 502-backend on every page, and the behaviour points squarely at the storage/infrastructure layer rather than my app.

What I'm seeing:

Even fully static requests are catastrophically slow. Requests for /favicon.ico and /apple-touch-icon.png (which don't run any of my application code) take 70–290 seconds to return a simple redirect. A static icon can't be made that slow by app code — this looks like file reads hanging on the storage backend. OSError: write error on essentially every request in the error log, right after each slow request. Workers won't recycle. The server log repeatedly shows worker N ... is taking too much time to die ... NO MERCY !!! across many PIDs, for minutes at a time — consistent with workers stuck in uninterruptible I/O. The Reload button fails with "There was a problem", and touch-ing the WSGI file doesn't recover the app either. Meanwhile the app itself is healthy: it imports fine, CPU usage is 0%, and disk is only ~8% used. So this isn't CPU throttling, a disk-quota issue, or a code problem on my side. I've verified locally that the exact deployed code serves /login and the DB pages in ~0.25s with no hangs, so I'm confident the cause is the platform, not my application.

Could staff please confirm whether the storage-cluster issue from this thread is still ongoing as of 27 July, and ideally hard-restart affected web apps' uWSGI instances? A status page or proactive notification during these incidents would also really help — right now the dashboard itself is intermittently unreachable, which makes it hard to even diagnose or reload.

Thanks!

We would like to apologise for the repeated instances of downtime and slow performance on PythonAnywhere over the last few weeks.

They have been the result of multiple issues that we have been dealing with. Each of them had an underlying cause that we were investigating. We believe that we have identified the causes of all of the issues and we are working on durable fixes for them. Unfortunately, there may still be some unexpected downtime as we try to apply the fixes that we have in time.

Thank you for your patience and understanding

So this latest downtime was planned. But the only place this was announced was X.com? Really? I try not to use X.com if possible.

Could you not notify customers via a more professional communication route? You have my email, after all. Or at least put something on my pythonanywhere dashboard.

For the most recent (planned) downtime, I got an email announcement yesterday, 27 Jul 2026 03:17:21 -0700 (PDT), with subject "[PythonAnywhere] Scheduled downtime notification: Tuesday 28th July 2026". I did have to opt-in to downtime notifications from my Account > Email page to get that. (I also avoid X.com as much as I can.)

We post via email, X and bluesky. Please double check your email settings.

Thanks.

Seems odd that downtime email alerts are "opt in" but hints and tips via email is "opt out". Changed now.

You might want to change your "service down" page. It only mentions service updates on x.com, not other platforms.