The mission is to keep the client’s service reachable through certificate renewals, including when the engineer who built the automation is absent. Before requesting a public key infrastructure (PKI) expert, ask one deciding question: can an authorized operator demonstrate that the renewed certificate reached every in-scope endpoint and that clients still trust the connection?
This guide applies Mission First and individual ownership from The Vitaver Playbook: Operating Doctrine & Craft. The staffing choices and pilot below are proposed practices, illustrated with a fictional service.
Trace the connection before writing the requisition
Imagine a manufacturer whose suppliers submit shipping notices through a portal. A renewal job reports success, but one load balancer still serves the old certificate. The certificate authority issued a replacement; the supplier’s connection can still fail. Treat issuance, deployment and application acceptance as separate checkpoints.
Have the application owner and platform lead draw the actual connection path. List public entry points, internal TLS connections and vendor-managed endpoints within the agreed scope. Record where TLS terminates, which service consumes the certificate, who controls installation and which user transaction depends on it. A network discovery result needs an owner’s review: it may miss an intermittently connected system or identify a device outside the engagement.
NIST SP 1800-16B’s certificate management guidance identifies an up-to-date inventory as a foundation and includes ownership and installation-location metadata. Use that reference to improve the working inventory. Keep private keys and credentials in the approved secret-management system, rather than attaching them to an inventory or staffing brief.
Buy the capability that removes the next obstacle
The hiring manager should classify a small sample of renewal problems before commissioning a large implementation. If nobody can identify the applications behind expiring certificates, begin with discovery and application-owner coordination. If the inventory is sound but deployment requires repeated manual intervention, prioritize automation. If certificates deploy but older clients reject them, obtain the application and trust-store knowledge needed to investigate compatibility.
| Responsibility | First useful output | Evidence to assess in a candidate |
|---|---|---|
| Platform automation engineer | A repeatable renewal and deployment path for the pilot service. | Can explain credential boundaries, deployment verification, failed steps and safe retries. |
| Certificate services or PKI specialist | Confirmed issuance, trust and replacement requirements for the chosen certificates. | Can distinguish a certificate problem from an incompatible trust chain or client configuration. |
| Application owner and receiving operator | A client transaction test and an actionable response to renewal failure. | Knows the application’s dependencies and can judge whether the service actually works. |
These are capabilities, not an automatic three-person external team. Existing staff may cover some of them. The client’s service owner accepts the operating result; the contractor owns the implementation assigned to them. Name a deputy for each time-sensitive client decision.
Give the agency a search-ready brief
Specify the first service, server and load-balancer technologies, certificate authority, existing renewal tooling and approved test environment. State whether the person must discover, design, script, deploy or support. Require exact product experience when the assignment depends on it; accept demonstrable adjacent experience when the client can support the remaining learning.
Confirm the actual site visits, maintenance windows, time-zone overlap, incident coverage and start prerequisites. An on-site need to reach an isolated appliance is different from a preference for someone nearby. Give the agency the budget boundary and expected engagement length, including client review and operator practice. Hiring an engineer before anyone can authorize deployment creates paid waiting rather than renewal capability.
The agency’s account owner should confirm candidate availability for those arrangements and distinguish reported experience from demonstrated work. The client should reserve one technical evaluator and prompt feedback against the agreed requirements. A second interview should resolve a different unanswered question.
Make the pilot prove the whole renewal
Choose one representative, lower-impact service with a named acceptance owner. Before work starts, agree the endpoints to verify, supported client configurations, an application transaction and conditions that stop expansion. Test issuance separately from the full deployment. Let’s Encrypt recommends its staging environment for integration testing; staging certificates are not a substitute for publicly trusted production certificates. Use the appropriate test facilities for the client’s actual issuer.
Ask the engineer to demonstrate a normal renewal and a controlled deployment failure in the test environment. For example, deny the test process permission to update one endpoint. The evidence should identify that endpoint, preserve the failure and notify the person who can act. A green issuance result must not overwrite a failed deployment result.
Verify what each endpoint actually serves using the intended hostname, then run the agreed client transaction with normal certificate verification enabled. Check the certificate identity, validity and trust chain from the relevant client path. Do not “fix” the test by turning off certificate validation. For a distributed service, record which endpoints and client paths were tested and which remain unverified.
Leave recovery time and a usable escalation
Renewal timing needs room for detection, repair and another attempt before expiry. Select it from the issuer’s current guidance, certificate lifetime and the client’s deployment constraints. Let’s Encrypt’s integration guide discusses renewal information, spreading renewals and retry behavior. Its settings should not be copied blindly to another issuer or certificate type.
NIST’s monitoring recommendations cover impending expiry and operating or configuration problems. For the pilot, agree who receives a failed renewal, a missed scheduled run and an endpoint still serving an old certificate. Set an escalation trigger that leaves the owner enough time to recover. Repeated alerts to an unattended mailbox do not establish coverage.
Document the approved recovery choice. An earlier certificate is usable only if still valid, trusted and not revoked or compromised. A suspected key compromise requires the security owner’s replacement and revocation process. Keep the emergency contact and authorized access route usable without relying on the failed portal.
Close with a client-led run
Before accepting the engagement, have the deputy operator follow the procedure, locate the pilot evidence and respond to the simulated failed deployment without the builder supplying missing steps. The service owner checks the transaction; the platform owner checks renewal and endpoint evidence. Record remaining manual systems with their owners and next review dates.
Review the inventory when an application moves, its hostname changes or its delivery platform changes. Expand automation only after the pilot exposes and resolves the material gaps. The purchased result is a renewal process the client can judge and operate, with exceptions visible early enough for someone to correct them.