Troubleshooting

Start with GET /status (Authorization: Bearer <ADMIN_TOKEN>): each post’s status and last_error. For more, the Worker’s logs: Cloudflare dashboard → your Worker → Logs, or pnpm exec wrangler tail. The sync’s output is in the build logs: dashboard → your Worker → Deployments (Workers Builds).

What you see	Cause	Fix
LINKEDIN_ACCESS_TOKEN is not set	The secret is missing	LinkedIn setup
LinkedIn 401 … invalid or expired	The token expired (60 days) or was revoked	Renew it
LinkedIn 403	The app lacks Share on LinkedIn, or the token was made without w_member_social	Add the product, generate a new token with the three scopes
LinkedIn 429	Too many requests today (150 per member)	Nothing: the retries continue; tomorrow resets it
media … was not found (did the sync run?)	The image is not in KV yet	Check the build logs; push again
The post stays scheduled after its time	The cron is not running	Dashboard → your Worker → Settings → Triggers: * * * * * must be there
missed	Pushed after its 10-minute window	Give it a new time and push
failed: the same text was already posted as …	Duplicate guard	Change the text, or delete the copy
unknown	LinkedIn did not answer clearly after the post was sent	Check your profile. If the post is not there, rename the file (a new name is a new post), set a new time, push
The push did not change anything	Workers Builds is not connected, or the sync stopped on an error	Build logs; Install
Nothing was synced, fix these first	A post file is invalid (date, time, status, media)	The message lists each file and what is wrong
… is over 25 MB	A video too big for KV	Make it smaller, or enable R2
does not have linkedin-scheduler v2's tables (build log)	The Worker is bound to a database made by something else: Cloudflare reused an existing database with the same name	Nothing was changed. Reinstall with a new database name, or point database_id in wrangler.jsonc at a new database
Posts show posted with a dry-run link, nothing on LinkedIn	DRY_RUN is "1"	Set "DRY_RUN": "0" in wrangler.jsonc and push (the dashboard value is overwritten by each push)
The first build failed right after the button	Usually a name already used on the account	Read the end of the build log; fix, then Retry build in the dashboard
Times are off by one or more hours	scheduler.json timezone	Set your IANA timezone (Europe/Paris, America/New_York…) and push
You scheduled “in 5 minutes” and it was already late	Your computer’s clock is wrong; the Worker uses Cloudflare’s	curl -sI https://www.cloudflare.com | grep -i '^date' gives the real time (GMT)