Domain health guide
SSL certificate expiry checklist
Certificate outages often appear when renewal, DNS, CDN, and origin ownership are not reviewed together. Use a repeatable checklist before a certificate reaches its renewal window.
Know who owns renewal
Document whether the certificate is managed by a hosting provider, CDN, load balancer, certificate service, or internal platform. Renewal failures are harder to resolve when the operator does not know which system holds the certificate lifecycle.
Also record the hostname coverage. A certificate that works for the apex domain may not cover www or another hostname used by customers.
Check HTTPS and DNS together
A certificate can be valid in one control plane but unavailable to visitors if DNS points to the wrong endpoint or a proxy routes traffic incorrectly. Confirm the public hostname, HTTPS reachability, and any redirect path after a change.
The current CloudSpex SSL checker reports certificate validation and HTTPS availability. It does not display expiry dates or days remaining, so use the certificate provider for expiry scheduling.
- Identify the certificate owner and renewal method.
- Confirm which public hostnames are covered.
- Check that DNS resolves to the intended public service.
- Verify HTTPS after renewal or proxy changes.
- Keep a monitoring and escalation owner for future changes.
Plan for change windows
Renewal is not only a certificate task. It can involve DNS validation, a CDN setting, an origin deployment, or a proxy. Schedule changes where they can be checked promptly.
If users report browser warnings, capture the hostname and time before changing settings. That makes it easier to distinguish a certificate problem from a DNS or routing problem.
What a public check cannot confirm
A public check can assess the response currently reachable for a hostname. It cannot confirm your renewal calendar, internal certificate inventory, private key handling, or every client-specific path.
Use it as a post-change signal, not as a substitute for certificate ownership and renewal management.
Why expiry is harder to manage than it used to be
Certificate lifetimes keep getting shorter. Publicly trusted certificates were once valid for several years, then for a maximum of 398 days. Certificates from Let's Encrypt are valid for ninety days. The industry body that sets the rules for certificate authorities has adopted a schedule that reduces the maximum further in steps, to 47 days by 2029.
The practical meaning is that renewal by calendar reminder no longer works. A certificate that must be replaced every few weeks has to be renewed by automation, and automation fails quietly. The task has moved from remembering to renew to noticing when the automatic renewal did not happen.
Build the inventory first
You cannot monitor a certificate you do not know about. Most expiry incidents involve a host that was not on anyone's list.
- The main website, with and without www.
- Subdomains that serve their own certificate: shop, app, api, portal, status, docs.
- Mail hosts: the names in the MX records and the hosts used for IMAP and SMTP submission.
- The mta-sts host, if the domain publishes an MTA-STS policy.
- Services on non-standard ports, such as admin panels and VPN gateways.
- Certificates held by a CDN, load balancer or hosting platform on your behalf, including the one between the CDN and your origin server.
Reading a certificate from the outside
The only certificate that matters is the one a client is actually served. Read it from the network, not from a file on the server. Three fields answer most questions: the validity period, the list of names the certificate covers, and the issuer.
openssl s_client -connect example.com:443 -servername example.com </dev/null 2>/dev/null \
| openssl x509 -noout -dates -subject -issuer -ext subjectAltName
notBefore=Sep 2 00:00:00 2026 GMT
notAfter=Dec 1 23:59:59 2026 GMT
subject=CN = example.com
X509v3 Subject Alternative Name:
DNS:example.com, DNS:www.example.comProblems that are not expiry
A certificate can be well within its dates and still produce a browser warning. These failures are checked by the same connection and are worth watching together.
- Name mismatch: the hostname is not in the certificate's list of names. Common after adding a subdomain or serving a site from a shared default virtual host.
- Incomplete chain: the server sends its own certificate without the intermediate. Desktop browsers often recover; mobile apps, API clients and mail servers usually do not.
- Wrong certificate served: the server holds the renewed certificate but was not reloaded, or one node behind a load balancer was missed.
- Untrusted issuer: a self-signed or internal certificate exposed on a public name.
- Revoked certificate: the authority withdrew it, for instance after a key was reported compromised.
How automatic renewal fails
ACME clients renew a certificate by proving control of the domain, either by serving a file over HTTP or by creating a DNS record. Either proof can stop working because of a change made for an unrelated reason.
A redirect or firewall rule starts blocking the validation path. The domain moves to a different DNS provider and the API token the client used no longer applies. A CAA record is added that does not name the certificate authority. The server is rebuilt and the scheduled job is not carried over. The certificate authority's rate limit is hit after repeated failed attempts. In every one of these cases the old certificate keeps working until the day it expires, so nothing looks wrong.
CAA records
A CAA record lists which certificate authorities may issue for a domain. Authorities must check it before issuing. It is a useful control, and a frequent cause of renewal failure when a team changes authority, or adopts a hosting platform that uses a different one, without updating the record.
example.com. CAA 0 issue "letsencrypt.org"
example.com. CAA 0 issue "digicert.com"
example.com. CAA 0 iodef "mailto:security@example.com"Alert thresholds that leave time to act
The right threshold depends on how the certificate is renewed. For a certificate renewed automatically thirty days before expiry, an alert at fourteen days remaining means renewal has already been failing for two weeks, which is exactly the signal you want. For a certificate renewed by hand, start earlier.
- Thirty days: a notice for manually renewed certificates, in time to raise a purchase or change request.
- Fourteen days: a warning for everything. Automated renewal should have succeeded by now.
- Seven days: urgent. Assign a named person.
- Three days and one day: escalate outside the usual channel.
- Expired: an incident, not an alert.
Who receives the alert
An alert is only as useful as its recipient. Expiry warnings sent to the address of a former employee, or to a shared mailbox nobody opens, are a recurring cause of outages. Send alerts to a monitored team address or channel, name a person responsible for each group of hosts, and record how each certificate is renewed and by which system, so that whoever receives the alert knows what to do.
After a renewal
A renewal is finished when clients see the new certificate, not when the tool reports success.
- Read the served certificate from outside and confirm the new expiry date.
- Confirm the list of names still includes every hostname that uses it.
- Confirm the chain is complete.
- Check every node behind a load balancer, not only the one you reached first.
- If a CDN sits in front, check both the CDN's certificate and the one on the origin.
When a certificate has already expired
Renew or reissue first and investigate afterwards. If automation is failing and cannot be fixed quickly, issue a certificate manually to restore service. Do not tell users to click through the warning; it trains them to ignore a protection that matters. Once service is back, find out why renewal failed and why no alert reached anyone, because the second question is usually the more important one.
Certificates outside the website
Web browsers show a certificate error immediately, so website certificates tend to be noticed. Other services fail less visibly. A mail server with an expired certificate may keep receiving mail from senders that do not verify certificates and lose it from those that do, including every sender honouring an MTA-STS policy in enforce mode. An API endpoint with an expired certificate breaks integrations and mobile apps while the website looks fine. A certificate used for a single sign-on or payment callback can stop logins or orders without any page showing an error.
Give these hosts the same monitoring as the website, and note for each one which port and protocol to check. Mail servers are commonly checked on port 25 with STARTTLS and on the submission and IMAP ports; the certificate on each can differ.
Frequently asked questions
Does CloudSpex show days until expiry?
No. The current public checker reports certificate validation and HTTPS availability, not an expiry countdown.
Can HTTPS work while the certificate still has a problem?
Yes. HTTPS reachability and certificate validation are related but distinct public signals.
Does the checker modify certificates?
No. It performs a limited, read-only public check.
How long before expiry should a certificate be renewed?
Automated clients typically renew with about a third of the lifetime remaining, which is thirty days for a ninety-day certificate. Manually renewed certificates should be started at least thirty days ahead.
Does a wildcard certificate cover every subdomain?
It covers one level. A certificate for *.example.com covers shop.example.com but not example.com itself and not a.b.example.com, unless those names are listed separately.
Why does the site work in a browser but fail for an API client?
Usually an incomplete chain. Browsers can fetch or cache a missing intermediate certificate; most other clients require the server to send it.
Use public checks as a starting point
CloudSpex reads limited public-facing signals. It does not change DNS, website, certificate, or email settings. Confirm the responsible provider and test relevant changes before applying a remediation.
Editorial details
Author and reviewer: CloudSpex. Sources: public technical standards and provider documentation relevant to the topic. Last reviewed: November 10, 2026. We update guides when the product behavior or the underlying guidance changes.