Does anyone know why this error happens? <@UK4HNFL...
# aws
r
Does anyone know why this error happens? @limited-rainbow-51650 or anyone from the Pulumi team or community can you help me?
l
Let me loop in one of our engineers.
@rough-jordan-15935 in the meantime, can you re-run with full verbose logging? https://www.pulumi.com/docs/support/troubleshooting/#verbose-logging
r
sure, will do 🙂
e
Qucik check, is this using Pulumi Cloud or s3/blob/file storage?
r
pulumi cloud
e
Are you behind a proxy server or any other networking systems that might be effecting calls to the serivce?
r
nope, no proxy 😕
e
Is this consistent?
r
this happens randomly
it happened yesterday night, and resolved after 1 hour
and it happened again today
and resolved after i updated pulumo
e
sigh always makes things harder 😞
r
but i'm not sure if this is fixed because i upgraded pulumi or if its a random fix
i was in 3.77 something before i upgraded. now i'm using this
e
This might be a server bug, but I'd need someone from our service team to investigate that. It might make sense for us to add a retry for these calls if they hit these errors though. Can you raise an issue about this at https://github.com/pulumi/pulumi/issues? We can look at adding the retry, and also loop the service team in there.
r
sure, is there any format that you have to create an issue?
e
If you hit "New Issue" one of the choices is "Bug Report", just follow the sections in that best you can.
r
done 🙂
hope i was descriptive enough
e
It's fine, thanks.
r
awesome
out of curiosity, how many resources would you recommend per stack @echoing-dinner-19531?
e
Generally whatever makes sense for your design, I don't think we really have a recommendation for what's a good size as it varies on some many other factors. Probably not thousands though because you'll start hitting some scaling issues, but some users do still do that.
r
does increasing the number of resources impact pulumi deploying the diff only?
because we have two stacks -
api
and
api-base
api
has very little resources and updates only the diff
api-base
has close to 600 resources and seems to update almost 20 to 30 resources per deployment that werent updated too
e
It impacts how long it takes to run the program, and there's some O(N^2) behaviour in saving state which means state saving takes longer and longer as you get more resources. Resources showing unrelated diffs is generally not because of the amount of resources, but how that provider is handling diffs for that resource type. You should generally report spurious diffs so we can check them out.
r
how do i report such a diff related issue?
Hey @echoing-dinner-19531, it wasn't a random fix. the error is back again
e
how do i report such a diff related issue?
Just raise an issue on the repo for the provider that gave the bad diff. Include what's changed in your code, what the diff is showing, and explain why you think it's wrong.
it wasn't a random fix. the error is back again
Yeh I would have been surprised if the the update had fully fixed it, this isn't an area of code we've touched recently. Hopefully some service team can take a look at this today and comment.
r
that is great to hear
🙂
@echoing-dinner-19531 discovered something else, i ran the deployment on codebuild and it worked. however, when i'm running it on the local development PC, it runs into an error. i tried switching networks and switching laptops, though this did not seem to do anything