Part 1 of an “Uneventful by design” blog series
At 2:00 a.m., a digital certificate expires.
It is a small file. Most people at the agency will never see it, and almost nobody outside the operations team knows it exists. But it is what allows one system to trust another. When it lapses, that trust breaks.
The application server that depends on it starts refusing connections. A middleware queue backs up. By the morning shift, a pharmacy order that should have moved to fulfillment overnight is still sitting in the queue. Somewhere, a Veteran checks the mail for a refill that is now a day late.
The fix takes 10 minutes. Finding the cause takes four hours. Explaining it takes a week.
Now replay the same night.
How certificate lifecycle management changes the game
Six weeks before the expiration date, an automated discovery scan flags the certificate. It sits in a central inventory alongside every other certificate in the environment, each one identifying its owner, the systems that depend on it and its renewal date.
Thirty days out, renewal runs on schedule. The new certificate is installed during a planned window, and an automated check confirms that every dependent service accepts it. The result is logged and reported.
At 2:00 a.m., nothing unexpected happens. The pharmacy order moves on time. The refill arrives on time. Nobody is paged, and nobody needs to know.
That second night is not luck. It is the product of an operating model.
Meeting the certificate management challenge
Certificate management used to be an annual chore. That era is ending.
Under the Certification Authority Browser Forum's schedule, the maximum lifespan of a public Transport Layer Security (TLS) certificate steps down to 47 days by 2029. A team that renewed a certificate once a year will soon renew it eight or nine times. Multiply that across every web front end, API and integration point in an agency the size of the Department of Veterans Affairs, and renewal volume grows by an order of magnitude.
Federal policy is raising the bar on visibility at the same time. Binding Operational Directive (BOD) 23-01 from the Cybersecurity and Infrastructure Security Agency (CISA) expects agencies to know what assets they run. OMB M-23-02 asks for an inventory of cryptographic systems in preparation for post-quantum cryptography. Spreadsheets and calendar reminders will not keep pace with any of this.
Certificate management proven at enterprise scale
The challenge is not unique to government. It appears wherever hundreds of interconnected systems must keep trusting one another.
For more than a decade, CGI has managed critical application services for Volkswagen Group, one of the world's largest automakers, with about 684,000 employees and vehicles sold in more than 150 countries. Volkswagen's application estate spans hundreds of systems across brands and geographies. Before consolidation, many different providers managed those systems, and the result was predictable: fragmented operations, uneven performance and limited visibility.
Bringing that delivery under one governed model, with shared standards and joint oversight, is what made consistent performance possible. The lesson carries directly to federal environments. Routine work scales only when one accountable model can see the whole estate, and certificates are the textbook example.
The principles at work
The difference between an outage and a non-event rarely comes down to what happens at the moment a certificate expires. More often, it reflects decisions made weeks or months earlier about visibility, automation and operational discipline.
Three principles matter most:
- Know everything you run. You cannot renew what you cannot see. Continuous discovery and a centralized certificate inventory create the visibility needed to manage certificates before they become a risk.
- Automate the predictable. Renewal, installation and verification are repeatable activities. Automating them reduces dependence on manual effort and frees people to focus on the exceptions that require judgment.
- Report the quiet. The absence of incidents is valuable operational data. Visibility into successful renewals, completed automations and avoided disruptions shows whether the operating model is working as intended.
These principles are not unique to certificate management. They hold wherever reliability depends on managing large numbers of routine tasks without disrupting critical services. The organizations that avoid certificate-related incidents are rarely reacting faster when something expires. They have built the visibility and repeatability that keep the expiration from becoming an incident in the first place.
Three questions worth asking
These strengths and weaknesses often stay invisible right up until a certificate expires. An outage does not create inventory gaps, manual processes or blind spots. It reveals them. Before certificate lifecycles get even shorter, three questions show whether the foundation for reliable certificate management is already in place:
- Could we identify every certificate in our environment, along with its owner and dependent systems, today? This reveals whether visibility is a strength or a vulnerability. Certificates that are unknown or poorly understood are often discovered only when they affect a service.
- Which renewals still depend on manual intervention? This identifies where operational risk is concentrated. As lifecycles shorten, repetitive manual steps become increasingly difficult to scale reliably.
- How far in advance do the right people know that a certificate requires action? Reliability depends on time to act. Effective notification and clear ownership turn renewals into planned work instead of emergency response.
The purpose of these questions is not to judge whether a program is perfect. It is to determine whether certificate management runs as a deliberate process or relies on individual effort and good timing.
Uneventful by design
Veterans should never feel the effects of an expired certificate. With complete inventory, automated renewal and reporting that shows what was prevented, they won't. CGI brings the same discipline proven across one of the world's largest commercial application estates to the federal systems Veterans depend on.
Next in the series: How rapid vulnerability remediation and steady service levels work together.