Back to blog
Threat Intel
Phishing Forensics

The Chaos of Equal-Priority MX: Debugging Proofpoint and Mimecast

Running two secure email gateways in parallel doesn't double your security, it creates a split-brain routing nightmare that threat actors ruthlessly exploit.

MailSleuth Research
Email Security Team
November 5, 20267 min read
Two glowing network pipes routing packets side-by-side, creating a chaotic vortex of data.

Project managers love a soft landing. When tasked with migrating a ten thousand seat enterprise from an old Secure Email Gateway to a new one, the immediate instinct is to avoid flipping a hard switch. Instead, someone proposes setting both the legacy system and the new platform to equal MX priorities. The logic seems sound on a whiteboard. Let the traffic distribute naturally, test the rules on both sides without dropping legitimate communications, and gently sunset the old infrastructure over a few weeks.

This is an architectural trap.

Running a proofpoint mimecast parallel mx configuration with equal priorities does not give you a redundant, smooth transition. It fractures your ingress routing, splits your quarantine boundaries, and turns standard incident response into a miserable scavenger hunt. DNS round-robin was designed for basic load distribution, not for stateful security policy enforcement. Here is exactly how this split-brain architecture falls apart in production and leaves your tenant wide open to evasion.

The Illusion of Load Balanced Security

When you publish multiple MX records with the same preference value, RFC 5321 states that the sending SMTP client must shuffle them and pick one at random. If that connection fails, it tries the next. Under normal internet weather conditions, those connections succeed. Your inbound mail flow splits right down the middle, governed by the unpredictable whims of tens of thousands of disparate sending Mail Transfer Agents.

You no longer have a single source of truth for email ingress. Half of your mail hits the Mimecast clusters, and half hits the Proofpoint instances. Both SEGs must now route downward to your Google Workspace or Microsoft 365 tenant. Your tenant receives mail from two entirely different sets of IP spaces, both of which must be aggressively whitelisted to bypass native Microsoft or Google filtering mechanisms.

Doubling the Ingress Attack Surface

You have just doubled your attack surface for tenant-level ingress. If an attacker discovers your Microsoft 365 direct routing hostname, they can attempt to spoof the SEGs directly. You lock this down using inbound connector IP restrictions. With two SEGs active, your connector must accept traffic from the massive IP blocks of both vendors. If a misconfiguration occurs on either SEG, or if an attacker compromises a shared tenant on either infrastructure, the blast radius spans your entire organization.

Policy Drift and Inconsistent Verdicts

Security gateways are highly stateful and deeply opinionated. Proofpoint Targeted Attack Protection and Mimecast Targeted Threat Protection do not score emails identically. They parse MIME structures differently. They normalize headers using different logic. They have different thresholds for what constitutes a Business Email Compromise attempt based on domain age, lookalike heuristics, and historical sender reputation.

When a threat actor launches a campaign against your domain, they rarely send just one message. They spray hundreds of targets across your organization. In a parallel MX setup, half of those malicious payloads hit the legacy SEG and half hit the new one.

Keeping rulesets perfectly synchronized across two different vendor architectures is technically impossible. A custom regex blocklist in Mimecast might not translate exactly to a Proofpoint custom rule. One platform might quarantine a message based on a failed DMARC alignment, while the other might softfail it due to a local ARC validation exception you forgot to mirror. A single phishing campaign yields two completely different outcomes depending on which MX record the sending MTA picked from the DNS response.

Dissecting a Split-Brain Phishing Campaign

Consider a credential harvesting campaign originating from a newly registered domain hosting a malicious PDF. The attacker blasts fifty executives. Twenty-five messages route through Mimecast. Twenty-five route through Proofpoint.

The Legacy Quarantine

Mimecast processes the first batch. Its URL rewriting engine inspects the links inside the PDF, but the newly registered domain has not triggered its specific threat intelligence feed yet. However, Mimecast enforces a strict policy on authentication. The attacker messed up their DKIM signature body hash during transit. Mimecast rejects the message at the gateway level. The SOC sees a nice, clean drop.

The New Gateway Delivery

Proofpoint catches the second batch. Its parsing engine successfully validates a secondary DKIM signature that Mimecast ignored, passing authentication. It scans the PDF and detonates it in a sandbox, but the payload has a logic bomb that sleeps for an hour. Proofpoint lets the message through to the inbox.

Authentication-Results: mx.proofpoint.com; spf=pass smtp.mailfrom=attacker.com; dkim=pass header.d=attacker.com; dmarc=pass action=none header.from=attacker.com;

Your SOC now has a crisis. Users report the phish. Analysts query Mimecast, see the blocks, and assume the campaign was successfully mitigated globally. They close the ticket. Meanwhile, twenty-five executives have the credential harvester sitting unread in their Outlook clients because the analysts failed to check the secondary gateway logs.

The Operational Toll on Incident Response

This fractured visibility breaks standard operating procedures. When a user reports a suspicious email, the first step for a SOC analyst is tracking the message ID through the gateway logs. In a split-architecture, the analyst must execute queries in two distinct administrative consoles. If the message hit the legacy SEG, but the analyst searches the new SEG, they get a false negative on the log search and waste critical triage minutes.

Automated ingestion into a Security Information and Event Management platform becomes a parsing nightmare. You feed two distinct syslog formats for the exact same entity types into your SIEM. Correlating a single threat actor's campaign requires building complex KQL or SPL queries that join Proofpoint message tracking logs with Mimecast MTA logs, accounting for differing timestamp formats and verdict naming conventions.

Releasing quarantined messages is equally miserable for end users. They receive two different digest emails from two different platforms. They authenticate to two different portals to find a missing invoice from a vendor. If a vendor complains about bounce backs, your messaging administrators have to trace the ingress route manually to figure out which gateway threw the SMTP 550 error.

The Right Way to Execute a SEG Migration

The desire to test a new gateway before a hard cutover is valid. The method of using equal priority MX records is fundamentally flawed. Defenders must utilize a staggered routing architecture that preserves a single source of truth for internet ingress.

Daisy-Chaining the Gateways

Point your primary, Priority 10 MX record exclusively at the new SEG. Configure the new SEG to route all processed, clean mail down to the legacy SEG, which then routes to the tenant. This daisy-chain approach allows you to run the new platform in observation or tag-only mode. The legacy platform remains the final authority on blocking. You can observe exactly how the new platform would have scored the email without risking split-brain delivery.

Once you validate the new gateway's ruleset, change its routing to bypass the legacy gateway and deliver straight to Microsoft 365 or Google Workspace. You then demote the legacy MX record to Priority 20 or remove it entirely. This guarantees that every single inbound message follows the exact same hop path.

If you must run side-by-side for departmental testing, do not touch the primary domain MX records. Provision a dedicated testing subdomain. Route that subdomain exclusively to the new SEG and forward copies of production traffic there for analysis. Threat detection requires absolute consistency. Your architecture must guarantee that an attack on your domain faces a single, unified defensive perimeter.

The takeaway

A secure email gateway migration is an inherently high-wire act. Attempting to soften the risk with equal-priority MX records only obfuscates your visibility and introduces critical operational failures. Threat actors do not wait for your change advisory board to approve a final cutover. They probe your defenses continuously, looking for exactly the kind of policy drift that parallel gateways create. Stop relying on DNS to balance your security posture. Force all traffic through a primary inspection point, daisy-chain your routing if you need sequential validation, and enforce strict inbound connector IP restrictions at the tenant level.

When you are debugging complex header chains and trying to determine exactly which gateway rewrote a URL or broke a DKIM signature, running a message through MailSleuth.AI can untangle the hop sequence in seconds. But no tool can fix a fundamentally broken routing topology. Pick a primary gateway, commit to the cutover, and pull the plug on the legacy system.

#email-security#mx-records#dns-routing#proofpoint#mimecast#soc-operations
MailSleuth Research
Email Security Team

We dissect phishing campaigns and email infrastructure so you don't have to.