- Key Takeaways
- Real-World Case Study: ResOnline's Certificate Expiration Outage, September 2025
- What Causes Certificate Outages in Enterprise Environments
- How Certificate Failures Disrupt Services
- Why Manual PKI and HSM Health Monitoring Falls Short
- What Continuous Monitoring Should Cover
- How Continuous Monitoring Helps Detect Outages Earlier
- How EC's PKI Prism Supports Continuous Monitoring
- A Practical Scenario: Catching an Expiring Certificate Before an Outage
- Conclusion
- FAQ
Many certificate outages begin with an expiration date that nobody was watching. A certificate reaches the end of its validity period, a Certificate Revocation List (CRL) is not republished, or a signing component quietly stops responding. Applications that depend on that trust can then fail, often without an obvious error pointing to the underlying Public Key Infrastructure (PKI).
These failures are often predictable, which is what makes them frustrating. Every certificate carries its expiry date, and every CRL carries a nextUpdate value. The information needed to anticipate many of these outages exists long before they occur. What is often missing is a reliable way to watch it, interpret it, and act on it in time.
This blog explains why certificate outages keep happening, what weaknesses in manual tracking allow them, and how continuous monitoring with Encryption Consulting’s (EC) PKI Prism, which monitors PKI and Hardware Security Module (HSM) health, helps teams detect problems earlier.Â
Key Takeaways
- Many certificate outages come from missed expiration dates and unobserved dependencies, not from weaknesses in the cryptography itself.
- Manual tracking through spreadsheets and reminders gives limited visibility between inspections and depends heavily on individual attention, which can delay fault detection.
- Continuous checks across the Certificate Authority (CA), its CRLs, and its signing dependencies can turn an emergency into a planned task, provided alerts reach the right owner.Â
- EC’s PKI Prism helps teams move from reactive troubleshooting toward proactive monitoring through recurring health checks, timely alerts, and centralized reporting.
Real-World Case Study: ResOnline’s Certificate Expiration Outage, September 2025
On September 6, 2025, ResOnline experienced a critical service outage after an Amazon Web Services (AWS) certificate failed to renew automatically and expired. The ResOnline operator console became inaccessible, and customers could not check availability or complete online bookings. ResOnline’s own incident report records the outage as starting at 10:00 a.m. AEST and being fully resolved by 11:40 a.m., about 1 hour and 40 minutes of disruption. Its booking gadgets and operator console on the resonline.com domain were affected, while distribution to online travel agencies was not.
The report notes that automated alarms did fire, but only after the certificate had already expired. That detection gap is what the rest of this article addresses.
In its follow-up actions, ResOnline planned to move its remaining core domains from email-based to DNS-based certificate validation and to improve monitoring so that expiring certificates are flagged early.
The incident illustrates why automated renewal alone is not sufficient. Early-warning alerts can give teams time to act before expiration and may reduce the likelihood of disruption.
What Causes Certificate Outages in Enterprise Environments
The ResOnline incident highlights a recurring problem: certificate outages often occur not because the warning signs are unavailable, but because they are not detected or acted on early enough. In enterprise environments, certificates are distributed across servers, applications, and infrastructure, making it difficult to maintain consistent oversight.
Several factors contribute to these outages:
- Certificate sprawl: Certificates are requested by many teams across load balancers, applications, devices, containers, and appliances. Without a complete inventory, some certificates are never tracked at all.
- Unclear ownership: The team or individual who requested a certificate may no longer be responsible for it, and the renewal notice may still go to a mailbox nobody reads.
- Expiry is only one failure mode: A certificate can be valid and still fail validation. A CRL that is not republished before its nextUpdate time, an unreachable CRL distribution point, or an Online Certificate Status Protocol (OCSP) responder that cannot sign responses can all disrupt applications that check revocation status.
- CA dependencies: An issuing CA may remain operational at the service level while being unable to perform signing operations because its HSM is unavailable. Depending on the CA configuration, this can interrupt certificate issuance or CRL generation and publication.
- Renewal failures: Automated renewal can fail because of validation issues, configuration errors, or missing dependencies.
- Configuration drift: A template change, a permission change, or a modified publication path can break issuance or validation, and the cause may not be noticed until services are affected.
- Shorter certificate lifetimes: Under CA/Browser Forum Ballot SC-081v3, the maximum validity of publicly trusted Transport Layer Security (TLS) certificates is 200 days for certificates issued from March 15, 2026, falling to 100 days from March 15, 2027 and 47 days from March 15, 2029. As validity periods shrink, each certificate needs renewal more often, and a process that worked at one renewal per year may struggle at that frequency.Â
These risks become harder to manage when monitoring relies on periodic manual checks rather than ongoing visibility into the infrastructure.
How Certificate Failures Disrupt Services
Certificates establish trust between applications, servers, users, and services. When a certificate expires, becomes invalid, or cannot be validated, dependent systems may reject connections or prevent cryptographic operations from completing. The resulting impact depends on how the certificate is used and how the application handles validation failures.
Depending on where the failure occurs, the disruption can appear in several forms:
- Website and API outages: Expired or untrusted TLS certificates can cause browsers and API clients to reject secure connections, making services inaccessible.
- Authentication failures: Expired, revoked, or untrusted client certificates can prevent users, devices, and applications from authenticating successfully.
- Application and service disruptions: Internal services that rely on mutual TLS (mTLS) may fail to communicate when certificate validation or trust-chain checks fail.Â
- Certificate issuance and renewal delays: CA service failures can prevent new certificates from being issued or existing certificates from being renewed on time.
- Cryptographic signing failures: HSM connectivity or authentication issues can prevent a CA from signing certificates and CRLs.
- Certificate revocation and validation issues: Failed CRL publication or unavailable OCSP services can disrupt certificate validation when relying systems require revocation checks.
- Business process interruptions: Certificate failures affecting payment systems, booking platforms, integrations, or automated workflows can delay transactions and interrupt critical operations.
These failures can extend beyond the component where the issue originates. A single certificate or infrastructure dependency may support multiple applications, increasing the potential impact. Understanding these dependencies helps teams identify which components require monitoring and where an undetected failure could disrupt business services.
Why Manual PKI and HSM Health Monitoring Falls Short
In many environments, certificate and CA health is tracked through expiry spreadsheets, calendar reminders, and periodic manual checks of CA services, CRLs, and HSM status. This can work at small scale, but it carries predictable weaknesses:
- Manual inventories go stale: A spreadsheet can become outdated as certificates are issued, renewed, replaced, or revoked.
- Reminders depend on individuals: An alert sent to one person may go unseen when that person is on leave or has changed teams.
- Discovery happens through failure: Users or applications can report the problem first, and the PKI team begins diagnosis under pressure.
- Diagnosis is slow: Finding the cause of a failure can require checking the CA service, CRL publication, event logs, and HSM status, often across several consoles.
- Knowledge is concentrated: Details of certificate dependencies, such as which applications rely on which CA or CRL, may sit with only a few administrators.
- Evidence is scattered: After an incident, reconstructing the timeline from separate logs can take considerable effort.
These weaknesses are not a reflection of administrator skill. They arise because periodic manual checks cannot reliably track time-bound conditions, such as certificate expiry and CRL nextUpdate times, or health changes that can occur between manual inspections.
What Continuous Monitoring Should Cover
Effective monitoring should look beyond certificate expiry dates and follow the full chain of trust:
| Monitoring area | What to check |
|---|---|
| CA service availability | Whether the issuing CA service is running and responsive, and whether it can complete issuance requests. |
| CA certificate and chain validity | How far away the expiry of the CA certificate and its parent chain is, since an expired CA certificate invalidates every certificate issued beneath it. |
| CRL validity and publication | Whether the base CRL and any delta CRLs are current, how far away their nextUpdate times are, and whether the CRL distribution point locations are reachable by relying systems. |
| OCSP responder health | Whether the responder is answering requests and whether its signing certificate and revocation data remain valid, for environments that rely on OCSP. |
| CA event logs | Whether errors and warnings, such as failed CRL publication or certificate services faults, point to a developing problem. |
| Signing dependencies | Whether the HSM or key storage backing the CA remains reachable and able to perform signing operations. |
| Alert routing | Whether warnings reach the responsible PKI or application team through channels they already use, with enough lead time to act. |
Monitoring these layers together matters because a failure in one, such as a signing fault, can surface later in another, such as a stale CRL. Tracking them as a connected chain, rather than as separate checks, helps teams find the originating fault sooner.
How Continuous Monitoring Helps Detect Outages Earlier
Continuous monitoring reduces the gaps between periodic checks by regularly assessing PKI components and alerting teams when issues require attention. Instead of relying solely on an engineer to discover a problem during a routine inspection, it provides ongoing visibility into the health of supported components.
Effective monitoring should focus on three areas:
- Early detection: Identify approaching certificate expirations, CRLs nearing their nextUpdate time, and supported CA or HSM health issues before they escalate.
- Timely alerts: Notify responsible teams when monitored conditions cross configured thresholds or indicate a potential failure.
- Consistent reporting: Maintain scheduled health reports and historical records to support troubleshooting, operational reviews, and incident investigation.
Monitoring supports earlier detection but does not guarantee that every outage can be prevented. Its value lies in reducing detection delays and giving teams more time to investigate and respond before a failure affects business services.
How EC’s PKI Prism Supports Continuous Monitoring
EC’s PKI Prism runs inside Claude Desktop and monitors Microsoft Active Directory Certificate Services (AD CS) along with supported HSMs, including Entrust nShield, Thales Luna Network HSM, and Thales Data Protection on Demand (DPoD), to help PKI teams detect and investigate CA and signing-path health issues earlier.Â
- Service availability checks: PKI Prism checks CA service availability at configurable intervals, which can be as short as 15 seconds, so a stopped or unresponsive CA can be flagged sooner.
- CRL status monitoring: CRL status is tracked against its nextUpdate time, so a CRL nearing expiry can be identified while there is still time to publish a new one.
- Configuration drift detection: Changes to monitored CA configuration compared with an expected baseline can be flagged, so unplanned changes, such as a modified publication path or CRL setting, can be reviewed before they break issuance or validation.
- CA chain and revocation service checks: PKI Prism also tracks the expiry of the CA certificate and its parent chain, the status of delta CRLs, the reachability of CRL distribution point locations, and the health of the OCSP responder, so problems in the wider revocation path can be identified alongside the CA itself.
- Log review: CA event information, including Windows Event Log entries, is collected so errors and warnings, such as failed CRL publication or certificate services faults, can be reviewed in one place alongside the other monitored events.
- AI-assisted investigation: Administrators can ask questions about their PKI in plain language through the Claude Desktop chat, such as which CRL is closest to expiry or what errors were logged recently, and Claude responds based on the monitoring data PKI Prism has collected. Responses should be treated as supporting evidence for the administrator to confirm.Â
- Alert routing: Alerts can be sent to Microsoft Teams, and tickets can be created in Jira, so issues are tracked to resolution by the responsible team.
- Reports: A detailed report covering CA service status, CRL status, collected events, alerts, and HSM health indicators can be generated and exported as a PDF, which supports review meetings, audits, operational reviews, and post-incident analysis.
- HSM health and performance: For supported HSMs, PKI Prism tracks availability and the performance and capacity indicators each platform exposes, such as CPU, memory, temperature, and cryptographic throughput on nShield and storage utilization on Luna, so a developing signing-path issue can be flagged earlier, before it affects certificate or CRL issuance.
- HSM resilience and logs: PKI Prism monitors nShield Security World status, module availability, and quorum configuration, along with Luna high availability (HA) group membership and health, and collects HSM logs, including nShield hardserver logs, Luna client logs, and Luna Network HSM appliance logs, so errors can be reviewed alongside CA events.Â
- Read-only design: PKI Prism is designed to be read-only. It collects and reports health information without making changes to the CA or HSM, and any corrective action is performed by the administrator.
Together, these capabilities give PKI teams a consolidated view of CA and HSM health, with alerts and reports that support earlier investigation. This helps teams act on warning signs while there is still time to plan the fix, rather than after services are affected.
A Practical Scenario: Catching an Expiring Certificate Before an Outage
An enterprise uses a TLS certificate to secure access to an internal application. The certificate is due to expire in three days.
With manual monitoring: The PKI team tracks certificate expiration dates in a spreadsheet and reviews it periodically, for example once a month. The administrator responsible for the renewal is on leave for a week, and no alert is generated by the spreadsheet. Unless someone opens it and checks the date, the approaching expiry goes unnoticed, and the renewal may be forgotten.
When the certificate expires, users encounter certificate warnings or connection failures, and administrators must investigate, renew the certificate, and restore service under pressure.
With EC’s PKI Prism: The team configures alert windows based on its own requirements, such as 60, 30, 7, and 2 days before expiry. Notifications are sent to the team’s shared channel as the certificate enters each window, so the renewal has already been flagged several times before three days remain.
If it is still pending, the next alert at the 2-day window notifies the team again, giving them another opportunity to plan and complete it. If the certificate remains unrenewed 24 hours before expiry, a final alert marks the urgency so the team can escalate before the expiration occurs.
By combining recurring alerts with clear ownership and a reliable renewal process, the team can address certificate expirations proactively rather than responding after an outage occurs.
Conclusion
Certificate outages are rarely surprising in hindsight. The dates were often available in advance, the dependencies could usually be traced, and warning signs may have appeared in logs before users noticed anything. What differs between organizations is whether those signals are watched on an ongoing basis and routed to someone who can act.
With public TLS certificate lifetimes shortening on a published schedule, manual tracking may become harder to sustain. Continuous monitoring of the CA, its CRLs, and its signing dependencies offers a practical way to find problems while there is still time to plan the fix.
EC’s PKI Prism is designed to support that approach. The goal is not to promise zero outages, but to detect warning signs earlier and reduce the likelihood that preventable issues escalate into service disruptions.
FAQ
What are the common causes of a certificate outage?
Common causes include expired certificates, failed renewals, unclear ownership, unavailable CRLs or OCSP services, CA service failures, and HSM connectivity problems.
How does continuous monitoring differ from manual checks?
Manual checks provide a snapshot of system health at a particular time. Continuous monitoring performs recurring checks, identifies configured warning conditions, and can alert teams between manual inspections.
Can monitoring prevent every certificate outage?
No. Monitoring improves detection and response but cannot eliminate every failure. Prevention also depends on reliable renewal and deployment processes, resilient infrastructure, and tested recovery procedures.
How can an expired CRL cause an outage?
Applications that require revocation checking may reject certificates when the CRL has passed its nextUpdate time and no newer CRL is available. The impact varies by application and by how it handles revocation failures.Â
Does EC’s PKI Prism make changes to the CA or HSM?
No. PKI Prism is designed to be read-only. Only explicitly permitted monitoring commands are executed, with controls intended to prevent write or state-changing operations.
- Key Takeaways
- Real-World Case Study: ResOnline's Certificate Expiration Outage, September 2025
- What Causes Certificate Outages in Enterprise Environments
- How Certificate Failures Disrupt Services
- Why Manual PKI and HSM Health Monitoring Falls Short
- What Continuous Monitoring Should Cover
- How Continuous Monitoring Helps Detect Outages Earlier
- How EC's PKI Prism Supports Continuous Monitoring
- A Practical Scenario: Catching an Expiring Certificate Before an Outage
- Conclusion
- FAQ
