Troubleshooting
Most Canopy issues fall into a handful of buckets. Use this page as a flowchart.
”My scheduled post didn’t publish”
Walk through these in order:
- Is the LinkedIn account connected? — Settings → LinkedIn. Red pill = reconnect required.
- Was publishing paused? — Settings → Pause publishing.
- Did the post status flip to FAILED? — My Posts → Scheduled tab. If you see a red pill, click for the error message.
- Was the scheduled time in the past? — should publish on the next scheduler tick (within ~1 minute).
- Was your account in a billing-failed grace period? — see Plans & billing → Failed payments.
If none of the above, email support@banyan.app with the post URL inside Canopy + the timestamp of the expected publish.
Common publish errors
| Error code | Meaning | Fix |
|---|---|---|
LINKEDIN_TOKEN_EXPIRED | OAuth token expired | Settings → LinkedIn → Reconnect |
LINKEDIN_PERMISSION_DENIED | Token missing w_member_social scope | Reconnect — re-authorize with full scopes |
LINKEDIN_RATE_LIMIT | Too many publishes in a short window | Auto-retries on next tick |
MEDIA_UPLOAD_FAILED | Image / carousel render didn’t upload to LinkedIn | Re-publish the post; if persistent, check the media file size is under 10 MB |
CONTENT_TOO_LONG | Post body over LinkedIn’s 3,000-char limit | Trim and re-publish |
CONTENT_REJECTED | LinkedIn rejected the content (rare; trust & safety) | Read the message, edit, re-publish |
NO_LINKEDIN_ACCOUNT | Post has no linkedinAccountId | Open the post, pick an account, save, retry |
”AI features aren’t working”
- Quota exhausted — Settings → AI → Monthly AI credits used. Reset on billing day, or upgrade plan.
- Provider error — Anthropic or OpenAI returned an error. Check the Anthropic / OpenAI status page.
- Auto-fallback failing — both providers are down (rare). Wait it out.
”Content Agent run failed”
Open the agent → Recent runs → the failed row shows the error. Most common causes:
| Error | Cause | Fix |
|---|---|---|
no_sources | No URL or keyword sources added | Add at least one source on the agent’s Notebook |
all_sources_failed | Every URL came back as FAILED | Check the URLs are publicly reachable |
model_no_usable_posts | Model returned no parseable JSON | Click Run-now again; if persistent, simplify tone notes |
quota_exceeded | Monthly AI credits exhausted | Upgrade plan or wait for reset |
”My calendar shows no slots”
You haven’t configured a weekly schedule. Settings → Posting Schedule → add at least one day-of-week + time pair.
”Analytics aren’t updating”
LinkedIn API delay. Force a refresh from the post detail (⋯ → Refresh analytics). For brand-new posts, wait 24–48h — LinkedIn sometimes hides the numbers initially.
”I see duplicate posts on LinkedIn”
Almost certainly:
- You scheduled the same draft twice (once via Queue, once manually).
- You clicked Publish now while a scheduled version was pending.
Canopy itself dedupes at every layer; duplicates are config issues, not bugs. Delete one of them on linkedin.com directly.
When to email support
Email support@banyan.app if:
- An error code isn’t in this page.
- A post is stuck in
SCHEDULEDpast its time without flipping toPUBLISHEDorFAILED. - Analytics for a post haven’t moved in >7 days.
Always include:
- Your user ID (top of Settings → Account).
- The post URL inside Canopy (and LinkedIn URL if it published).
- Approximate timestamp of the issue.
We can pull worker logs and tell you exactly what happened.