You set up Cloudflare email forwarding, sent yourself a test message, and it arrived. Two weeks later a client asks why you never replied to an invoice question. You check the Cloudflare dashboard: the rule is there, the destination looks right, and the activity log shows the message came in and went nowhere. No bounce, no error, no trace in your inbox. That is a silent drop, and it is the most common failure mode in Cloudflare Email Routing.
The mechanics explain why. Cloudflare Email Routing forwards email for domains using Cloudflare DNS, which means mail hits Cloudflare’s MX records first and then gets relayed to your destination mailbox. Every hop in that chain is a place where a message can be accepted and discarded: an unverified destination address, a provider that filters Cloudflare’s forwarding IPs, a rule that matches but has no valid target, or an authentication check at the receiving end that decides the forward looks like spoofing. None of these produce a bounce back to the sender, because Cloudflare already accepted the message. The sender sees a successful delivery, and you see nothing.
This checklist exists to catch those failures before your customers do. It walks through each handoff point in the pipeline — destination verification, provider-side filtering, SPF and DKIM alignment, rule design against documented Cloudflare Email Routing limits, and ongoing log review — with the specific status values, record types, and test procedures that tell you whether mail is actually landing. Run it once after setup, then again after any DNS change, nameserver delegation, or destination mailbox migration. The whole pass takes about fifteen minutes and saves you from the slow discovery that a forwarding rule has been quietly eating mail for months.
How Cloudflare Email Routing Works and Where Forwarded Mail Gets Lost
Cloudflare Email Routing is a free forwarding layer for domains that use Cloudflare DNS. You point the domain’s MX records at Cloudflare, publish its SPF and DKIM records, and Cloudflare accepts mail for your addresses and relays it to destination mailboxes you control, typically a Gmail inbox. The path looks simple on a diagram. In practice there are five handoffs, and a message can disappear at any of them without a single bounce reaching the sender.
Here is the pipeline. The sending server resolves your MX records, which point at Cloudflare’s mail exchangers. Cloudflare accepts the SMTP connection, checks the recipient against your routing rules, and re-injects the message toward the destination address. Your destination provider then runs its own spam and authentication checks before the message lands in the inbox. Each arrow in that chain is a place where mail can stop.
The Five Handoff Points Where Mail Disappears
- MX resolution. If nameservers are not delegated to Cloudflare, the Email Routing records never take effect and mail goes to whatever MX the old DNS provider still advertises. That mailbox may be dead, and the sender gets a delayed bounce at best.
- Rule matching. Cloudflare evaluates routing rules in order. A custom address with no matching rule is rejected outright unless a catch-all exists. With a catch-all set to «drop,» unmatched mail is discarded and the SMTP conversation still ends with a 250 acceptance, so the sender believes delivery succeeded.
- Destination verification. Destination addresses must be verified before rules activate. Until you click the link in the verification email, rules pointing at that address stay inert and messages are dropped, not queued.
- Relay to the destination provider. Cloudflare forwards from its own infrastructure. Gmail, Outlook, and corporate gateways frequently score forwarded mail on the reputation of the forwarding IP and the authentication results of the original sender, and can reject or silently filter it.
- Final delivery at the destination. The message can pass every check and still land in spam, a filtered label, or a quarantine that never notifies you.
The uncomfortable part is that Cloudflare Email Routing reports most of these outcomes only in its activity log, and some not at all. There is no queue you can inspect and no retry you can trigger. When a message is dropped, the log shows a status of «dropped» with a reason code, but nothing is sent to the sender and nothing is sent to you.
Treat the pipeline as five separate checkpoints rather than one service. The rest of this checklist walks through each one in order, from destination verification to reading the activity log after every DNS change.
The Silent Drop Checklist: Verifying Your Destination Address
Cloudflare Email Routing will accept a message for your domain and then discard it if the destination address was never verified. No bounce goes back to the sender. The message does not land in a spam folder. It simply disappears, and the only evidence is a status line in the activity log that most people never open. This is the single most common cause of «my forwarding stopped working» reports, and it is almost always self-inflicted: someone added a destination address in the Cloudflare dashboard, assumed the rule went live, and moved on.
The mechanics are strict. Destination addresses must be verified before rules activate. When you add an address, Cloudflare sends a confirmation email to that mailbox with a link you have to click. Until that click happens, any routing rule pointing at the address exists in the dashboard but does not deliver. If you are forwarding email to Gmail, check the spam and promotions tabs for the verification message, because Google’s filters sometimes treat automated sender-verification mail as bulk. If the link expired, resend it from the dashboard rather than deleting and recreating the address.
Work through this checklist in order. Each step takes under a minute and rules out a whole class of silent drops.
- Confirm the address shows as verified. In the Cloudflare dashboard, open Email Routing and check the destination addresses list. Every entry should carry a verified status. Any entry still pending is a black hole for mail.
- Match each rule to a verified destination. Each routing rule supports one destination address. A rule pointing at an unverified or deleted address will drop, so audit every custom address and the catch-all target individually.
- Send a real test from outside your domain. Use a personal Gmail, Outlook, or work account that is not hosted on Cloudflare. Mail sent from within the same domain can take a different path and will not prove the inbound pipeline works.
- Wait, then check the activity log. Open the Email Routing activity log and find your test message. A delivered status means the chain is intact. A dropped status means Cloudflare received the message and chose not to forward it, which points at verification, a rule mismatch, or a provider rejection.
- Test the catch-all separately. The catch-all rule catches addresses not matched by other rules. Send to a random address at your domain, such as a string you have never used, and confirm it appears in the log with a delivered status.
- Re-test after any edit. Changing a rule, removing a destination, or reordering rules can silently break delivery. Treat every configuration change as a reason to run the external test again.
One detail worth knowing: Cloudflare Email Routing does not support sending or replying from the domain. That means you cannot verify the pipeline by replying to a forwarded message and watching it come back. The test has to originate from an external account every time. Keep a note of which external mailbox you use for testing so results stay comparable between runs.
If the activity log shows delivered but the message is not in the inbox, the drop happened downstream at your mail provider, not inside Cloudflare. That is a different problem, and the next section covers how to diagnose it.
Checking Whether Your Destination Provider Blocks Cloudflare’s Forwarding IPs
Cloudflare’s outbound forwarding servers send a lot of mail, and receiving providers know it. Gmail, Outlook, and most enterprise filters score incoming messages partly on the reputation of the sending IP and the envelope domain. When Cloudflare hands your message to Gmail, the envelope sender is a Cloudflare-controlled address, not yours. If that infrastructure is throttled, greylisted, or outright blocked by the destination provider, Cloudflare has already accepted the message and considers its job done. The activity log will show the forward as delivered, while Gmail quietly files it to spam or rejects it after the SMTP handshake with a 5xx that Cloudflare records but does not bounce back to the original sender.
Spotting a block in the activity log
Open the Email Routing activity log in the Cloudflare dashboard and filter by the destination address. Each entry shows a status: delivered, dropped, or an error. A «dropped» status on a rule whose destination is verified and whose address is spelled correctly is the strongest signal of a downstream rejection. Look at the timestamp distribution as well. If messages from one sender domain fail consistently while others pass, the problem is content or sender reputation, not the destination. If everything to one mailbox fails at once, suspect the provider’s filtering of Cloudflare’s forwarding IPs.
Confirm from the other side. Send a test message from an external account (not another address on the same domain) to the routed address, then check both the Cloudflare log and the Gmail spam and trash folders. A message that appears in spam is a filtering decision, not a drop, and it tells you the pipeline works. A message that never arrives anywhere, with a delivered status in Cloudflare, points to a provider-level block or a silent discard in the recipient’s tenant.
Workarounds when the provider blocks forwarding
- Allowlist the Cloudflare forwarding domain in the destination provider’s admin console. In Google Workspace, add it under Gmail settings, then Compliance, then Blocked senders, or as an approved sender in the spam filter. Consumer Gmail has no such control, so this only works for managed accounts.
- Route to a different mailbox. A second destination address on another provider (Fastmail, Proton, a self-hosted server) tells you in minutes whether the block is provider-specific. You can keep the original rule and add the alternate as a separate routing rule, since each routing rule supports one destination address.
- Forward through a Worker instead. Worker email processing lets you rewrite headers, wrap the message, or send it through a transactional provider with a known-good sending reputation. This is heavier to maintain but bypasses the shared forwarding pool entirely.
- Check the message size. The maximum message size is 25 MiB. Anything larger is rejected before it ever reaches the destination, and that rejection looks identical to a provider block in the log.
Document which workaround you applied and the date. Provider filtering changes without notice, and a fix that held last quarter may not hold now.
Configuring SPF, DKIM, and DMARC for Forwarded Messages
Forwarding changes the envelope sender. When Cloudflare Email Routing relays a message, the SMTP MAIL FROM becomes a Cloudflare-controlled address, while the visible From header stays as the original sender. SPF is checked against the envelope, not the header, so the receiving provider evaluates Cloudflare’s infrastructure and not the author’s domain. That mismatch is exactly what DMARC alignment catches. If the original domain publishes p=reject and the forwarded copy fails alignment, Gmail and Microsoft 365 will quarantine or discard it, and the activity log shows the handoff as delivered while the message never reaches the inbox.
Cloudflare handles the sending side for you. When you delegate nameservers to Cloudflare, Email Routing provisions the MX records for inbound mail and adds SPF and DKIM records so its outbound relay is authenticated. Do not edit or delete those records. If your zone already had a TXT record at the apex with an SPF policy, check that Cloudflare’s include is present and that you are not exceeding the 10 DNS lookup limit that SPF enforces. A lookup overflow returns permerror, which most receivers treat as a fail.
DKIM is the stronger lever. The signature is computed over the message body and selected headers, so it survives the envelope rewrite as long as nothing modifies the content in transit. Cloudflare signs relayed mail from its own selector. Confirm the selector record exists in the zone and resolves with a public key, and confirm the destination provider verifies it. In Gmail, open the message, choose Show original, and check that SPF reads pass or neutral and DKIM reads pass with the domain aligned to the From header. A pass on DKIM with alignment is what keeps a forwarded message out of spam even when SPF breaks.
At the destination domain, publish your own records rather than relying on the source. If you forward to a Gmail mailbox, you cannot change Google’s inbound policy, so the practical fixes are:
- Keep DMARC at
p=nonewith reporting enabled on any domain you control that forwards mail, and read the aggregate reports before tightening to quarantine or reject. - Add a DMARC record at the destination domain if you own it, and align DKIM with the From domain you actually send as.
- For domains you own that forward to Gmail, avoid strict alignment (
adkim=s,aspf=s) until you have confirmed forwarded mail passes DKIM.
ARC is worth knowing about. Some receivers honor the ARC chain that Cloudflare and intermediate hops add, which preserves the original authentication result across forwarding. Not every provider validates ARC, so treat it as a safety net, not a guarantee. Test with a real message from an external account after every DNS change and read the authentication results in the received headers before you conclude the records are correct.
Cloudflare Email Routing Limits and Rule Design That Prevent Drops
Most silent drops I have investigated in Cloudflare Email Routing came down to a rule that did exactly what it was told. The limits are published, but they are easy to skim past when you are setting up forwarding in a hurry. Here is what actually constrains a domain, and how rule design interacts with those constraints.
Documented Limits and What They Mean in Practice
Cloudflare Email Routing allows up to 200 routing rules per domain. That sounds generous until you build per-user aliases for a 60-person team with role addresses on top. Once you hit the ceiling, new rules cannot be created, and any address you forgot to cover falls through to whatever the catch-all does, which may be nothing.
Each routing rule supports exactly one destination address. A rule cannot fan out to three mailboxes. If a team wants an alias delivered to several people, you need either multiple rules for the same custom address or a destination-side group. The first option burns through your 200-rule budget fast; the second is usually the better design.
Destination addresses must be verified before rules activate. An unverified destination does not produce a bounce to the sender. The message is accepted at the MX and then discarded, which is the classic silent drop. Every destination address needs its confirmation link clicked, and re-verification after a provider change or a long period of inactivity is worth checking in the Cloudflare dashboard.
Maximum message size is 25 MiB. Anything larger is rejected at the edge, and depending on the sending server you may or may not get a useful NDR. If your business receives large attachments, PDFs from scanners, or CAD files, tell senders to use a link instead of an attachment.
Cloudflare Email Routing does not support sending or replying from the domain in its basic form. That matters for rule design because outbound replies from your Gmail account will carry the Gmail address, not the alias, and recipients may treat the mismatch as suspicious. Outbound sending lives under Cloudflare Email Service, which groups Email Routing and outbound sending together, so plan for it if you need alias-preserving replies.
Rule order is not cosmetic. The first matching rule wins, so a broad pattern placed above a specific one will swallow the specific address and send it to the wrong mailbox, or to a mailbox nobody reads.
The catch-all rule catches every address not matched by other rules. That is useful for typo tolerance and for aliases you have not documented, but it is also where mail disappears when the catch-all points to a stale destination or is set to a drop action. Audit it quarterly. A catch-all aimed at a departed employee’s inbox is a quiet data loss event.
Keep a written map of rules: custom address, destination, order, and purpose. When you add a rule, place specific addresses above wildcards, confirm the destination is verified, and send one external test message to prove the route works before you move on.
Ongoing Monitoring: Reading Activity Logs and Re-Testing After Changes
Forwarding that works today can stop working next month without a single configuration change on your side. A destination provider tightens its spam filters, a colleague deletes a mailbox, someone edits an SPF record during a domain migration. None of these trigger an alert to you. The only reliable signal is the Email Routing activity log in the Cloudflare dashboard, which records a status for every message that touches the pipeline.
Read it on a schedule. Once a week is enough for most domains; daily if you run catch-all rules that feed a shared inbox. Sort by status and look for three values: dropped, rejected, and delivered. Delivered means the handoff completed. Dropped almost always points back to an unverified destination address or a rule that matched and then failed to resolve a target. Rejected usually means the receiving server refused the connection, which is where provider-side blocking shows up. A message that appears in the log with no terminal status after several minutes deserves a manual resend to confirm it actually landed.
Pair the log review with scheduled external test sends. Internal tests prove nothing because they often bypass the MX path entirely. Send from a Gmail account, an Outlook account, and one address at a smaller provider you rarely use. Rotate the subject lines so they do not look identical across weeks, and include a short body rather than an empty message. Keep a note of the timestamp, then find that exact message in the activity log within a few minutes. If the log shows delivered but the inbox is empty, the loss happened after Cloudflare handed the message off, and you are looking at a provider filter rather than a routing problem.
Re-run the full checklist after any of these events:
- Nameserver delegation changes or a move to a new DNS host
- Edits to MX, SPF, or DKIM records at either the forwarding or destination domain
- Adding or removing a destination address, which resets its verified status until the confirmation link is clicked
- Changes to catch-all rules or the order of routing rules
- Migrations at the destination provider, such as a Gmail workspace move or a mailbox rename
- Any report of missing mail from a real sender, treated as a live incident rather than a curiosity
Keep a short log of your own: date, change made, test send result, log status. It takes two minutes and it turns a vague «mail seems flaky» into a timeline you can act on. Remember that Cloudflare Email Routing does not send or reply from your domain, so outbound tests must always originate from an external account. And because each routing rule supports one destination address, a rule edit that adds a second recipient silently replaces the first rather than duplicating delivery. Check the rule list after every edit, then send the test. That habit catches more silent drops than any single setting.
Put the checklist on a schedule instead of treating it as a one-time setup task. A workable routine looks like this: once a month, open the Cloudflare dashboard, go to Email Routing, and scan the activity log for any status other than delivered. Once a quarter, send a test message from an external account (a personal Gmail address works fine) to every destination address you rely on, including one address that only matches the catch-all rule, and confirm each one lands in the inbox rather than the spam folder. After any DNS change, nameserver move, or mailbox provider migration, run the full list again within 24 hours, because MX, SPF, and DKIM edits at either end can break forwarding in ways that produce no bounce at all.
Start today with the single highest-value step: verify every destination address and check the activity log for dropped messages from the past week. If the log shows drops with no obvious cause, work through the list in order, confirmation first, then provider filtering, then authentication records, then rule design. Keep a short note of what you changed and when, so the next test send has something to compare against. If your needs have outgrown Email Routing, for example you need outbound replies from the domain or programmatic handling of incoming mail, look at Cloudflare Email Service and Worker email processing before adding more forwarding rules on top of a setup that is already stretched. The goal is not a perfect configuration on day one; it is a forwarding chain you can prove works every time you check it.