Why prospects might fail and how to fix
When a LinkedIn or email step fails for a specific prospect, the campaign Prospects tab shows a per-prospect error pill explaining why. Most errors are temporary and resolve on the next scheduler tick — but a small set are permanent and need user action. Here is the full list:
- Invite failed: HTTP 404 — LinkedIn could not find the profile. Common causes: profile deleted by the user (most common), account banned by LinkedIn, member-id changed after a profile rebrand, or rare Unipile bridge cache lag.
What to do: A single 404 can be transient and the platform will retry. After 3 consecutive 404s, the campaign detail page shows an 'Invalid prospects' banner with a one-click Exclude button. Excluding a prospect flips it to status='stopped' with reason 'Excluded: profile inaccessible (HTTP 404)', cancels any pending send jobs, and frees the daily-invite quota slot. Excluding fires zero LinkedIn API calls — it is a database-only operation, so it is safe even if the underlying account is in cooldown.
- Account in cannot_resend_yet cooldown until <timestamp> — LinkedIn told Unipile your account hit its invite limit. Per Unipile's documentation, this is HTTP 422 and the platform's recommended response is to space requests at random intervals and respect LinkedIn's 80-100 invitations/day band (5/month for free accounts). WarmySender sets a 4-8h account-level cooldown so other prospects on the same account also wait — burning more sends during a throttle deepens the rejection.
What to do: WAIT. The cooldown auto-clears at the timestamp shown in the error. Do not pause and resume the campaign — that does not shorten the cooldown and adds noise. The platform takes GREATEST(existing, new) on every cooldown signal, so repeated throttles never shorten the deadline; only longer ones extend it. The job that hit the cooldown gets marked expired (terminal for that retry attempt), but the underlying enrollment is rescheduled to the cooldown end + jitter, and a fresh job mints automatically on the next tick.
- Daily LinkedIn limit reached / Weekly invite limit — You sent up to your configured daily/weekly cap, which is anchored to LinkedIn's 80-100/day band and your ramp week (1-4).
What to do: WAIT. The daily cap resets at midnight UTC; the weekly cap resets Monday 00:00 UTC. Scheduler resumes automatically.
- Outside sending window — The action would have fired outside your campaign's sending window (e.g. weekend, late evening).
What to do: WAIT. The job re-fires on the next valid window slot. If you want sends 24/7, edit the campaign and widen the window — but for LinkedIn, off-hour activity raises ban risk.
- linkedin_account_disconnected (campaign pause) — LinkedIn cookie expired or you logged out elsewhere.
What to do: RECONNECT. Settings → LinkedIn → Reconnect. The campaign auto-resumes within 5 minutes (one scheduler tick) once the connection is healthy.
- Permanently failed after 5 stuck — A processing safety valve fired. Usually caused by 5 consecutive transient blockers (disconnected account, daily-cap hit, sending-window starvation).
What to do: RECOVER. The campaign detail page shows a 'Stuck prospects' orange banner with a Recover button — click it once the underlying issue is resolved (account reconnected, capacity added, etc.).
For the full HTTP 404 troubleshooting and cooldown deep-dive (including Unipile's official guidance), see the 'Why prospects might fail and how to fix' section at /documentation#prospect-failures.