DKIM configuration is straightforward at deployment.
Generate a key pair, publish the public key in DNS,
configure the MTA to sign outbound messages with
the private key, test signature validation, declare
DKIM deployed. The configuration works correctly at
the moment of deployment and continues working
indefinitely without intervention. The operational
gap appears later when audit, security, or
compliance requirements expose the absence of
ongoing key management hygiene.
The first gap is rotation. M3AAWG Best Common
Practices document recommends regular rotation
(originally quarterly, updated to semi-annual in
2019) but most operations never rotate. The
original DKIM key from 2019 still signs messages in
2026 because nobody scheduled the rotation, nobody
owns the rotation procedure, and the rotation
tooling does not exist. Auditors reviewing SOC 2
CC6.7 or ISO 27001 A.8.24 controls find that the
organisation has DKIM configured but cannot
demonstrate key rotation as part of cryptographic
hygiene. The audit finding either downgrades the
related controls or requires remediation before
report issuance.
The second gap is selector strategy. Operations
typically deploy one DKIM selector covering all
outbound mail. The shared selector creates the
compromise blast radius problem (one stolen key
enables impersonation across every service), the
rotation coordination problem (rotation requires
synchronised reconfiguration), and the diagnostic
difficulty problem (DMARC failures cannot identify
the source service). The M3AAWG and SPFmonitor 2026
guides both document per-service selector isolation
as the operational pattern that addresses all
three. Adoption requires upfront design and
ongoing management which the Premium tier provides.
The third gap is the dual-selector cutover.
Rotation cannot be atomic because deferred messages
remain in transit. A message sent at 14:00 signed
with the old private key may not arrive at the
recipient until 14:23 or queue for hours of retry
windows. If the rotation happens at 14:15 and the
old public key gets deleted from DNS immediately,
the deferred message fails validation when it
arrives. Operations that attempt manual rotation
without dual-selector practice frequently break
deferred mail validation for the rotation window.
The Premium tier runs the dual-selector cutover
standard with 14-day overlap window covering all
plausible deferred mail scenarios.
The fourth gap is replay-attack monitoring. The
Valimail October 2025 continuous-protection report
and SPFmonitor 2026 guide document DKIM replay
attacks as the primary residual risk after key
rotation. An attacker who captures a legitimate
signed message from your domain can re-send the
message to other recipients later; the DKIM
signature still passes because the attacker has not
modified anything. The result: messages from your
domain getting delivered to recipients you never
intended, with the original DKIM signature still
validating. Operations without replay monitoring
discover replay activity only when recipients
complain or when domain reputation degrades.
Operations with replay monitoring detect the
pattern in TLS-RPT and external monitoring data
and can trigger emergency rotation before
reputation damage accumulates.
The fifth gap is emergency rotation capacity. Key
compromise scenarios (team member leaves with key
access, system breach affecting key storage,
replay attack detected, external disclosure
affecting cryptographic primitives) require
immediate rotation rather than waiting for the
scheduled cycle. Operations without rotation
tooling and procedure are not positioned to
execute emergency rotation when needed. The
Premium tier includes emergency rotation with
compressed timeline (1-4 hour cutover, 7-day
overlap window) at no per-event surcharge,
providing the operational capacity for incident
response rather than requiring it be developed
under pressure.