How to split big worker into smaller ones to minimize startup time?

Hi all! we've hit a limit recently "- Error: Script startup exceeded CPU time limit. [code: 10021]" when trying to deploy our API worker. https://github.com/AsyncStatus/asyncstatus/tree/main/apps/api. Deploy and commit that introduced this issue https://github.com/AsyncStatus/asyncstatus/actions/runs/17192217490 Deploy and commit that fixed this issue https://github.com/AsyncStatus/asyncstatus/actions/runs/17192500191 (but we would like to add more and more code so it's not really representative but you get the idea). You can see it does a lot of things in just one index.ts (apps/api/src/index.ts) like exporting workflows, durable objects, scheduled handlers, queues and normal handlers (we have API and webhook handlers too). Not sure how to ask this question but I guess to avoid startup time limit error we could split this one worker by the function? Meaning one worker for exporting workflows, one for durable objects, one for "normal" handlers, one for queues etc... Or maybe split by domain? So we'd have like status-update-api or organization-api apps? Or maybe we should just fill out the form to increase startup limit? But we're a small startup so we wouldn't be able to pay +$100 for this request. Also cpuprofile file generated by wrangler check startup is not really useful, it would be great to have more human friendly output. Thank you, Kacper

GitHub

feat: send linear activity · AsyncStatus/asyncstatus@fa4c2af

Async status updates for remote startups. Contribute to AsyncStatus/asyncstatus development by creating an account on GitHub.

GitHub

chore: remove unused handlers · AsyncStatus/asyncstatus@3db5bda

Async status updates for remote startups. Contribute to AsyncStatus/asyncstatus development by creating an account on GitHub.