I was simply trying to run code from the Anvil interface and was failing. And some server code did not seem to be replying as usual… wondering if Anvil was down (and if it is down how do you broadcast the problem?)
All seems fine right now … just was wondering for about 15 minutes…
I started getting server timeout errors this morning. Logs show I’m calling the server on the client side but never hit the first line of the server function on the server side. I’ve tried from the browser and from a server instance via a command line call.
I hate to assume its an Anvil issue rather than something on my end, but I’m out of ideas. Is this the right place to ask this?
It’s also a good place to use the Forum’s Search feature. You can often find an answer faster that way, than waiting for someone to reply.
In the past, this has happened to me when I inadvertently had an Uplink program running that provided an @anvil.server.callable of the same name. Since Uplink callables are prioritized over Server-side callables, my Uplink program was getting triggered, and that’s the one that timed out.
Thanks. I had checked the forum previously, but didn’t find anything that really seemed to fit. I don’t have any Uplink programs. It definitely seems like either something is blocking the server call, or I’m just not connecting to the server. I tried calling other functions from the server CLI and get the same non-response, so if that’s any indication, I don’t think its necessarily a function-name specific issue.
I am having a very odd issue where my server (AWS EC2) responds quickly for a particular @anvil.server.callable function when called from a client on my phone browser, while it will complete on the server but does not successfully pass data to a client on my laptop browser! No idea why.
How often are you getting a server timeout? (like what % of the time)
I sometimes get this with one of my apps, and I’m always running a VPN, so I also think that internet strength can be an issue.
We’ve been seeing increased load and traffic from apps on the Free plan. This is great (more people using Anvil!), but has caused some strain on the infrastructure we use for serving that traffic. If you have a Free account, you’ll probably have noticed your server code is slower than usual – sorry about that. We’re busy improving the infrastructure for serving Free users!
(Paid users shouldn’t have been affected – we keep resources for Free users segregated from paid users, for precisely this reason.)
2. Load balancing configuration error
As we were introducing our new infrastructure for free users, this exposed a a configuration error in our load-balancing configuration. One node within our cluster had lost the distinction between resources for serving free and paid users. This isn’t normally a problem – that node was previously not responsible for managing server code. However, as part of the changes we’re rolling out to improve server side performance, this node was now taking part in load-balancing. So for a little under 24h, this node was periodically reallocating paid users’ server code onto the (very heavily loaded) free-user infrastructure.
Current status
This configuration error has been identified and resolved as of mid-Tuesday. We apologise to the affected users.
We’re continuing to experience heavy load on our Free-user-facing infrastructure. We’re working to expand provision for our Free users – so if you’re on a Free account experiencing nasty bumps, hold tight!
Would there be somewhere a general “Anvil server health” page available where status for free servers and business (paid) servers have separate indicators?