Allowing notifications: how to keep important e-mails out of spam
A notification about an important e-mail in the security mailbox only helps if it reaches the inbox. Spam filters often take automated e-mails with urgency and a link for phishing. This guide shows how your IT allows the sender safely.
All notifications from CRA Response Center come from the address app@cra-response-center.com. To keep them out of the spam folder, your IT allows this sender once: in Microsoft 365 with a mail flow rule, in Google Workspace with a list of approved senders, in an upstream mail gateway with a permitted sender entry. Allow by authenticated sender, not by IP address. A test alert then shows whether the notification reaches the inbox.
Key points
- The sender of all notifications is app@cra-response-center.com, the domain is cra-response-center.com.
- The domain is protected with SPF, DKIM and DMARC (policy reject). An allow entry that requires passed authentication lets no forged e-mails through.
- Do not allow by IP address: the sending servers are shared with others, and their addresses can change.
- Everyone who is notified should also add the address to their own address book.
- After allowing the sender, send a test alert and check that the e-mail is in the inbox.
Why do notifications end up in spam?
A notification has exactly the features spam filters look for: it is automated, it names a deadline, it asks for quick action and it contains a personal link. Phishing e-mails look similar. On top of that, learning filters such as the junk filter in Thunderbird or Outlook compare every e-mail with those previously marked as spam. If a notification arrives in a language that is rare in the mailbox, the filter lacks experience and would rather sort it out.
Authenticity is usually not the problem. Every notification is signed with DKIM, the SPF record of the domain lists the permitted sending servers, and the DMARC policy of the domain asks recipients to reject e-mails that fail the checks. Allowing the sender only tells the filter: nothing from this authenticated sender should be sorted out.
| Where | Allow entry | Who sets it |
|---|---|---|
| Microsoft 365, Exchange Online | Mail flow rule: sender domain plus the check "dmarc=pass", bypass spam filtering | Administrators |
| Google Workspace | Spam setting with a list of approved senders, require sender authentication | Administrators |
| Gmail (single account) | Filter with "Never send it to Spam" | Each person |
| Outlook | Safe Senders list | Each person |
| Thunderbird | Address book and "Not Junk" | Each person |
| Mail gateway (e.g. Mimecast, Hornetsecurity, Proofpoint, Sophos) | Permitted sender for the address, with an SPF, DKIM or DMARC check | Administrators |
The labels come from the providers' help pages. Interfaces change, and individual words may differ in your version.
Microsoft 365 and Exchange Online
For allow entries, Microsoft first recommends the Tenant Allow/Block List. By default, however, Microsoft removes its entries for domains and addresses 45 days after they were last used, and a specific expiry date can be at most 30 days ahead. Notifications are rare, often months apart. A mail flow rule, which Microsoft describes as the second option, therefore works permanently. The steps according to Microsoft's help:
- Create a ruleIn the Exchange admin center, go to Mail flow, Rules and create a new rule.
- Limit the senderCondition "The sender domain is" with the value cra-response-center.com.
- Require authenticationSecond condition "A message header includes any of these words": header Authentication-Results, words dmarc=pass and dmarc=bestguesspass.
- Bypass spam filteringAction "Set the spam confidence level (SCL)" to "Bypass spam filtering". Microsoft also advises setting a custom header, such as X-ETR with a note about the rule, so the allow entry can be traced later.
What to watch for: Microsoft explicitly warns against building such a rule with the sender domain alone, because attackers could spoof the domain. The second condition with dmarc=pass prevents that. According to Microsoft, messages identified as malware or high confidence phishing are quarantined regardless of any allow entry. If Microsoft takes a notification for phishing, report it as clean in Submissions in the Microsoft Defender portal.
Official help: Create safe sender lists and Allow or block email using the Tenant Allow/Block List (Microsoft Learn).
Google Workspace and Gmail
In Google Workspace, administrators set the allow entry for the whole organisation. The steps according to Google Workspace help:
- Open the settingIn the Admin console, go to Apps, Google Workspace, Gmail, Spam, Phishing and Malware and select the top-level organisational unit.
- Create a ruleUnder Spam, click Configure or "Add another rule" and enter a name.
- List of approved sendersChoose "Bypass spam filters for messages from senders or domains in selected lists", create a new address list and add app@cra-response-center.com.
- Keep authentication onThe option that requires sender authentication stays on; according to Google it is set by default for new entries. Then save.
What to watch for: Google warns against approving senders without authentication. For a private Gmail account without administrators, each person creates a filter: open the search options in the search bar, enter app@cra-response-center.com under "From", click "Create filter" and tick "Never send it to Spam".
Official help: Customize spam filter settings and Allowlists, denylists, and approved senders (Google Workspace help).
Outlook and Thunderbird
Learning junk filters in a person's own mail program decide separately for each person. The administrators' allow entry does not change that. So everyone who is notified should add the sender once themselves:
- Outlook: add app@cra-response-center.com to the Safe Senders list, in the Junk Email options or by right-clicking a notification. The list only applies to that person's mailbox.
- Thunderbird: add the sender to the address book and, in the account settings under Junk Settings, set that mail from senders in this address book is not marked as junk. Mark a notification that was already sorted out as "Not Junk" so the filter learns from it.
- Other programs: almost every mail program has safe senders or a "never spam" rule. Adding the address is enough, an entry by domain is not needed here.
Mail gateways in front of the mailbox
Many companies run their own filter in front of Microsoft 365 or their mail server, such as Mimecast, Hornetsecurity, Proofpoint or Sophos. It can reject or hold e-mails before they reach the mailbox; then no setting in the mailbox helps. Create a permitted sender there (in Mimecast "Permitted Senders") for app@cra-response-center.com. An entry for the single address is narrower than one for the whole domain and is enough. If the filter offers to tie the entry to passed SPF, DKIM or DMARC checks, switch that on.
Mimecast itself advises against allowing whole domains or large IP ranges, because spoofed e-mails would otherwise get through. If the gateway holds e-mails in a queue, administrators see them there and can release them; the notification then arrives late, though.
Why not allow by IP address?
Many guides recommend allowing a sender's IP addresses. For notifications from CRA Response Center, that is the wrong way. The e-mails go out through sending servers in a data centre that other customers there use as well. Allowing these addresses would let their e-mails through unchecked too. The addresses can also change without notice, and the allow entry then silently stops working.
Microsoft writes that e-mail from sources in the IP Allow List skips spam filtering and sender authentication with SPF, DKIM and DMARC, and advises against ranges of shared infrastructure. Allowing by authenticated sender is narrower and stays correct when servers change. The servers permitted to send are always listed in the SPF record of the domain cra-response-center.com.
Common mistakes
- Allowing by domain only. A rule without a check also lets forged e-mails with this domain through. Always require dmarc=pass or sender authentication.
- Only administrators allowed the sender. The junk filter in Outlook or Thunderbird sorts it out anyway. Everyone who is notified adds the sender themselves.
- The allow entry expires. Entries in Microsoft's Tenant Allow/Block List disappear after 45 days without use. For rare e-mails, use a mail flow rule.
- The gateway was forgotten. An upstream filter rejects before the mailbox sees the e-mail. Allow the sender there too.
- No test. Only a test alert shows whether the notification reaches every person's inbox.
Frequently asked questions
Which senders do we need to allow?
One address: app@cra-response-center.com. All notifications, invitations and sign-in e-mails come from it. Where a rule is built by domain, it is the domain cra-response-center.com.
Is an allow entry a security risk?
An allow entry without a check is one, because senders can be forged. With the condition dmarc=pass or with sender authentication required, only e-mails that provably come from the domain get through. With its DMARC policy reject, the domain asks for all others to be rejected.
Do we need to allow the text messages as well?
No. Text messages do not pass through e-mail spam filters. Some phones sort text messages from unknown senders into a separate folder; in that case save the sender CRA as a contact on the phone.
How do we test whether it works?
In CRA Response Center, send a test alert under Settings, Test alert. Each person checks that the e-mail is in the inbox and not in the spam folder or a quarantine, and confirms the test alert.
Does this have anything to do with forwarding?
No, that is the other direction. Forwarding brings e-mails from your security mailbox to CRA Response Center; for that, your mail system must allow forwarding to external addresses. The allow entry here makes sure the notifications get back to your people.
Will a notification certainly arrive with an allow entry?
The allow entry removes the most common hurdle. Delivery by e-mail still depends on many parties, and Microsoft quarantines detected malware and high confidence phishing despite any allow entry. The security mailbox itself remains the source, and you keep reading it.
Further reading
- Setting up forwarding for the security mailboxWhere forwarding sits with common mail providers and what often makes it fail.
- Spam in the security mailboxWhy the security mailbox's spam filter should mark rather than reject.
- The first 24 hours after a vulnerability reportWhat to do after the notification.
Sources
- Create safe sender lists, Microsoft Learn, as of 25 October 2026.
- Allow or block email using the Tenant Allow/Block List, Microsoft Learn, as of 25 October 2026.
- Customize spam filter settings, Google Workspace help, as of 25 October 2026.
- Common pitfalls in Mimecast email security configurations, Mimecast, as of 25 October 2026.