Every email leaving Lumail goes through a shared sending system. Understanding it explains
why a campaign takes the time it takes, and why a workflow email sometimes arrives a few
minutes after the action that triggered it.

## Two priority lanes

| Lane        | What goes in it                                                                                       | Priority                         |
| ----------- | ----------------------------------------------------------------------------------------------------- | -------------------------------- |
| **Instant** | Transactional emails (login codes, notifications, API/SMTP single sends) and campaign **test emails** | Always sent first                |
| **Bulk**    | **Campaigns and workflow emails**                                                                     | Sent with the remaining capacity |

Whenever an instant email is waiting, bulk sending pauses until it goes out. Instant emails
are rare and tiny in volume, so this pause is measured in milliseconds — but it guarantees
that a password reset never sits behind a 50 000-recipient campaign.

## Workflow emails are queued like campaigns

Each email step in a workflow is backed by a campaign behind the scenes. That means:

- Workflow emails use the **same bulk lane** as regular campaigns.
- They join the **same sending queue for your organization**.
- They are **not prioritized** over campaigns, and campaigns are not prioritized over them.

Inside one organization, the queue is **first in, first out**: whatever was queued first is
sent first. If a large campaign starts sending at 10:00 and a subscriber triggers a workflow
email at 10:01, that workflow email is sent once the campaign ahead of it has finished
draining.

If you need a workflow email to go out immediately regardless of campaign volume, avoid
scheduling large campaigns during the windows where that workflow is most active.

## Fair share between organizations

Your queue is never blocked by another customer's campaign. Sending capacity is split
across all organizations that have emails waiting:

- Half of the available capacity is **reserved for paid plans**.
- The rest is open to every organization, free plans included.
- Within each pool, capacity is distributed by plan weight: **Business** organizations get
  more slots than **Premium**, which get more than **Free**.
- Organizations that were short-changed in one round are prioritized in the next one, so
  the split evens out over time.

Practically: if you are the only organization sending, you get the full throughput. If ten
organizations are sending at once, capacity is shared — and your plan determines your share.

## What this means for timing

- **Small campaigns and workflow emails** typically leave within seconds.
- **Large campaigns** are delivered progressively, at a steady, deliverability-friendly
  rate rather than in one burst. This is intentional: sending everything at once damages
  sender reputation.
- **Progress is visible on the campaign page** while sending, so you can follow how many
  recipients have been processed.
- Emails are never dropped when the queue is busy — they wait, and each recipient's state
  is tracked individually so a retry never produces a duplicate.

## When sending stops

Sending pauses for your organization in two cases:

- **Monthly quota reached** — campaigns are parked and workflow runs are paused. Both resume
  automatically at the start of the next period, or as soon as capacity frees up.
- **Plan suspension** — same behavior: nothing is lost, everything resumes once the account
  is active again.

## Related Documentation

- [Email Step](/docs/workflows/workflow-email-step) - Send emails from a workflow
- [Deliverability Score](/docs/features/deliverability-score) - Monitor your sender reputation
- [Subscriber Events](/docs/features/subscriber-events) - Track delivery, opens and clicks