How Do You Triage a Certificate Transparency Alert for Your Domain?
Published 2026-11-14 · Reviewed 2026-11-14 · 6 min read · By the Ingenious AI team
Is someone impersonating your domain right now?
Run the free DomainGuard scan: lookalike registrations, MX-armed variants, certificate-transparency and CAA checks for your domain, in one report.
Run my free DomainGuard scanThe short answer
Triage a Certificate Transparency alert by asking, in order: is the name ours, did we request this, and what else exists at that name? Certificates you or your vendors requested — renewals, new subdomains, a platform migration — explain most alerts. An alert is an incident when the name is a lookalike you do not own, when the subdomain is one nobody on your team recognises, or when the name is yours but no change ticket matches the issuance. Work those three questions within a day; a hostile certificate is often the last public signal before a phishing site goes live.
Why certificates are such a good tripwire
Almost every credible phishing site uses HTTPS, because browsers mark plain HTTP login pages as unsafe and users have learned to glance for the padlock. Getting that padlock means a certificate, and every publicly trusted certificate is published to CT logs at issuance — a global, timestamped, searchable record. The attacker cannot quietly prepare the final step of their infrastructure. They can register in the shadows, but the moment they make the site presentable, a log entry announces the exact hostname to anyone watching.
That is why CT monitoring sits at the center of DomainGuard's detection: it converts the attacker's own professionalism into your early warning. In the worked examples throughout these guides, the recurring pattern — clone found at certificate issuance, filed on before the mail run — is not luck. It is the structural consequence of watching the one feed the attacker cannot avoid feeding.
Question one and two: ours, and requested?
Start with the name itself. A certificate for your own domain or a subdomain of it goes down one path; a lookalike name — typo, homoglyph, combo, your brand inside someone else's domain — goes down another, and the lookalike path is an impersonation finding by definition. For your own names, ask whether the issuance matches a known action: certificate renewals are the bulk of all alerts (automated 90-day cycles mean they arrive constantly), followed by new services, vendor platforms you pointed subdomains at, and infrastructure migrations.
Reconcile against evidence, not memory: your certificate inventory, your DNS change log, your vendors' status notices. A 'probably us' that nobody can tie to a ticket is not a reconciliation. The uncomfortable middle case — your domain, a plausible subdomain, no matching change — must be treated as hostile until explained, because the ways an attacker gets a certificate for your real subdomain (a dangling CNAME takeover, a compromised DNS account) are precisely the incidents you most need to catch early.
Question three: what lives at that name?
For anything unexplained, look at the name itself — carefully, and from a safe environment. Resolve it, note the host and any CDN in front, and load the site in an isolated browser profile you do not use for real accounts. You are answering a short list: is something served, does it copy your brand or login, does it ask for credentials or payment, and is there mail infrastructure (MX) alongside? Screenshots with the address bar become your evidence pack if the answers are bad.
If the name is a lookalike, also check its siblings: attackers certificate in batches, and the same day's logs often reveal the campaign's other hostnames. DomainGuard's triage groups alerts by registration and infrastructure patterns for exactly this reason — one alert is a finding, but the cluster is the incident, and filings that name the cluster protect you against the attacker's reserve domains as well as the one you spotted.
The response ladder by verdict
Explained and ours: log it, update the certificate inventory, close the alert — and let the ease of that closure remind you why the inventory exists. Unexplained but inert (lookalike registered, certificate issued, nothing served yet): file the registrar report with the CT evidence, raise the name's watch level, and brief the team that a build is in progress. Live and cloning: the incident playbook runs at full speed — blocklists and host in the first hour, registrar in parallel, customer warning through unforgeable channels.
Timing discipline is what makes the ladder work. Same-day triage keeps you ahead of the mail run in most cases, because attackers typically certificate a day or more before distribution — testing, cloning, warming the domain. A weekly alert-review rhythm spends that gift. Route CT alerts for certificated lookalikes as same-day items, and let bare renewals on your own domain batch into the weekly digest where they belong.
Reducing alert noise without blinding yourself
Alert fatigue is the real enemy of a CT program: a busy estate can generate hundreds of legitimate issuance events a month, and a stream nobody reads is worse than none, because it feels like coverage. The cures are classification, not silence. Suppress by rule only what is provably routine — your known CAs, your known subdomain patterns, your vendors' certificate profiles — and review suppression rules quarterly, because a rule broad enough to hide your own platform's renewals might also hide a takeover wearing a similar shape.
CAA records shrink the unknown pile from the other side: with CAA published, only your named CAs may issue for your domain, so a certificate from any other CA is automatically a high-severity alert. That is the cleanest triage signal available in this entire field — one record turns a judgement call into a binary. Publish CAA, keep the inventory honest, and your CT queue becomes what it should be: quiet for weeks, and impossible to ignore when it speaks.
Five actions to take today
- Classify each CT alert by name: your domain, your subdomain, or a lookalike — lookalikes are findings by definition.
- Reconcile your own names against the certificate inventory, DNS changes, and vendor notices; 'probably us' is not a verdict.
- Inspect unexplained names safely: host, content, credential forms, MX records — and screenshot everything with the address bar.
- Apply the ladder: log the explained, file on the inert lookalike, run the full incident playbook on the live clone.
- Publish CAA records and keep suppression rules narrow, so routine renewals stay quiet and foreign-CA issuance is always loud.
Worked example: Brightpath Tutoring separates 112 alerts from 1 incident
Brightpath Tutoring, a fictional SaaS monitoring 3 domains, received 113 CT alerts in a quarter. Triage classified 108 as its own platform's automated renewals and vendor issuances, 4 as inert lookalike registrations worth registrar filings, and 1 — a certificate for a login-combo lookalike from a CA the company had never used — as an incident. That single alert, actioned the same morning, led to a login clone taken down before its first mail run. The team's total triage time for the quarter was under 3 hours, because classification, not volume, was the workload.
FAQ
We get renewal alerts constantly. Is something wrong?
No — short-lived certificates renew every 60 to 90 days per name, so a large estate generates steady alert volume. Classify renewals by known CA and name patterns into a digest, and keep only anomalies loud.
Can an attacker get a certificate for my real domain?
Only by controlling something of yours: a dangling subdomain target, a DNS account, or a validation channel. That is why a foreign certificate on your own name is a high-severity alert — it implies a foothold, not just imitation.
What if the alert is for a name that does not resolve?
Certificates precede deployment routinely. Log it, watch the name, and file the registrar report if it is a lookalike — you are seeing the attack during setup, which is the cheapest moment to act.
Does CAA stop rogue certificates completely?
CAA instructs compliant CAs not to issue outside your list, and major CAs must check it. It dramatically reduces unauthorised issuance for your domain — but does nothing for lookalike domains, which is why CT watching covers both.
How fast should we triage?
Same day for lookalikes and foreign issuances on your names; weekly digest pace is fine for routine renewals you can match to inventory automatically.