Sudden appearance of CPU time limit
Since today I have had the following error:
Error: Script startup exceeded CPU time limit. code: 10021]
I haven't changed my application code but now wrangler CLI or the Dashboard is no longer able to deploy.
I also tried to just update environment variables and redeploy from the dashboard but I still have the same problem. If i haven't change my code why am I unable to deploy?
Is this related to the issue on March 4? https://www.cloudflarestatus.com/incidents/3q7sg3tpp7qv
I tried deploying the app with "less code" and it worked by returning a single new Response() object. Then progressively increasing the code gets random deployments pushed.
I've also upgraded to the "Paid" plan to see if that resolves the issues if I am reaching some limit.
Did Cloudflare modify some CPU limits since today??
Please help 🙏🏻 this is affecting production release
5 Replies
Just to add: Upgrading to a "Paid" plan in workers didn't fix the issue
I have a custom domain bound to the worker application and I noticed that my simple deployment (just returning new Response('Hello, world!', { status: 200 }) ) don't update the worker or the domain.
Additional information: My application runs a GraphQL Apollo Server with the
"@as-integrations/cloudflare-workers": "^0.2.0", library
Is Cloudflare Workers ok??? I am now having these errors in the dashboard
Unrelated but harmless error, it's being escalated as we speak, just annoying visually right now
Edit: Should be fixed now
Sadly upgrading does not increase the startup time:
https://developers.cloudflare.com/workers/platform/limits/, it's 400ms regardless of plan.
You have too much trying to execute in global/taking up too much cpu time. I'm not aware of any recent changes with this, and that incidient you linked was changing the CPU Limit (not startup) for paid plans.
I'd check on what you're going in global scope/your dependencies.
Here I post an update I posted on #workers-discussions:
I have found and applied a solution that was moving an import inside the cloudflare workers handler, so it reduced the startup time. Reference example here: https://github.com/cloudflare/workers-sdk/issues/3096#issuecomment-1560883961
I am using apollo server on cloudflare workers with this library "@as-integrations/cloudflare-workers": "^0.2.0".
But what I found weird is that the dashboard didn't work either. Also, the deployment was delayed, and even changing to a different network didn't work.
I am testing it now and seems like it's working again, and the last deployment is now using the updated environment variables
Did Cloudflare change some limits?
GitHub
🐛 BUG: Startup script exceeded CPU time limit · Issue #3096 · cloud...
Which Cloudflare product(s) does this pertain to? Workers/Other What version of Wrangler are you using? 2.17.0 What operating system are you using? linux Describe the Bug I get this error from the ...
By the way, thanks for looking into this