Skip to content

Why your certificate renewed on schedule but still didn’t get there

Summary

Renewing a certificate does not confirm it has reached the system that needs it. An accurate inventory and independent verification of the certificate actually being served help close the gap between successful renewal and successful deployment, across internal and external services. The starting point is clear: discover what is deployed, catalogue it, and monitor it independently.

 

It's Monday morning and authentication is down. Not for one application. For everything behind it. Helpdesk tickets are stacking up before the coffee's finished brewing, and the incident channel fills with the same question from three different teams: didn't we renew that certificate last month?

They had. The change record closed on schedule. The certificate authority confirmed issuance. Every system involved in the renewal reported success. What nobody confirmed was whether the renewed certificate actually reached the server that needed it. It hadn't.

The reason is simple. Issuance and deployment get confirmed by the same systems that performed them. Nobody independently checks whether the certificate that was supposed to land on an endpoint is actually the one being served there, until it fails.

This isn't rare, and it isn't a story about any one team missing something. Security leaders are telling researchers directly that they're struggling to keep up: 42% admit they lack a unified approach to securing machine identities at all, according to CyberArk's 2025 State of Machine Identity Security Report. 72% have already experienced at least one certificate-related outage in the past year, and among those affected, roughly a third lost between $50,000 and $250,000, with nearly one in five losing more. Why this is accelerating, and what's actually driving it, is covered in the FAQ below.

This isn't only a public website problem. Internal certificates, on Active Directory, on database connections, on email services, fail exactly the same way, often with less visibility because nobody's watching as closely.

There's usually a second issue sitting underneath the first: certificates nobody catalogued in the first place. Discovery, done properly, tends to surface more than expected.

Our process - start with an accurate starting inventory

Discover → Catalogue → Deploy → Monitor

Most organisations already have tools handling Deploy. What's usually missing is independent confirmation at the Monitor stage, and an accurate starting inventory to monitor against.

How we start: At Avocado, we start with a short, no-cost conversation about your current certificate estate. If it's worth going further, we run a scoped Discovery engagement, cataloguing what's actually deployed across your environment, internal and external, and flagging anything unaccounted for. That baseline becomes the foundation for ongoing, independent verification.

FAQ: Why is certificate deployment getting worse?

Why are certificate-related outages becoming more frequent?

Volume and speed have both increased at once. Machine identities, certificates, keys, secrets, and tokens, are proliferating fast, driven by AI adoption, cloud-native architecture, and now agentic AI. Machine identities already outnumber human ones by a wide margin, and organisations expect up to 150% further growth over the next year. More certificates in circulation simply means more chances for one to be mishandled.

Isn't shorter certificate validity supposed to make this safer?

It closes one risk while opening another. Shorter lifespans reduce how long a compromised certificate stays valid, but they also mean far more frequent renewal cycles. 37% of security leaders already report difficulty keeping up with accelerated renewal and rotation requirements. Each additional renewal cycle is another opportunity for a deployment to be missed, not caught.

Does cloud infrastructure make this harder?

Yes, in a specific way. 74% of security leaders are concerned about managing identities in cloud-native, ephemeral workloads, where infrastructure spins up and down constantly. That makes it harder to maintain a stable, accurate inventory of what even needs a certificate at any given moment, let alone confirm each one is correctly deployed.

Why doesn't better automation solve this on its own?

Because automation and verification are different jobs, and most organisations only have the first one covered. Responsibility for certificates is typically split across security, development, and platform teams, with no single team accountable for confirming the end-to-end outcome. Automating the renewal step faster doesn't create anyone whose job is to check the result.

How much has this actually grown?

CyberArk's 2025 State of Machine Identity Security Report found 67% of organisations now experience certificate-related outages monthly and 45% weekly, up from 26% monthly and 12% weekly in 2022. That's roughly a 2.5x and nearly 4x increase respectively in just a few years, not a marginal shift.

Related content