Skip to main content
Video tutorial · 0:50

Reissue a certificate

Reissue generates a fresh private key and requests a new certificate from the same issuer with the same subject / SANs. It is the right choice when you suspect key compromise, when you want to change the key algorithm (e.g. RSA 2048 → ECDSA P-256), or when policy mandates periodic key rotation independent of the expiry-driven renewal cadence.

Not available for imported certificates

Reissue requires a CA to re-request from. Imported (bring-your-own) certificates have no issuing CA, so Re-issue is rejected for them — the refresh path is Re-import (upload a fresh certificate).

Reissue vs. renew​

RenewReissue
Triggered byExpiry window / schedulerExplicit operator action
Private keyReused by default — a new key only when the certificate's key-rotation policy is set to rotateAlways NEW
Subject / SANsSameSame
IssuerSameSame
Use caseScheduled lifecycleKey compromise, policy, algorithm change

The key is the important difference. Renewal reuses the existing private key unless the certificate opts into rotation, so waiting for a scheduled renewal after a suspected key compromise would leave the compromised key live. Reissue always generates a new key pair.

The other distinction is timing: reissue fires now, regardless of the renewal window and regardless of ARI hints.

After a key compromise, reissue — do not wait for renewal

Renewal keeps the same private key by default. If you suspect the key is exposed, reissue immediately (and revoke the old certificate), rather than relying on the next scheduled renewal to replace it.

When to use reissue​

  • Key compromise. Followed immediately by revocation of the old cert.
  • Algorithm upgrade. Moving from RSA 2048 to ECDSA P-256 fleet-wide; reissue with the new key type, then revoke the old on completion.
  • Deployment proof. Some change-management workflows require a fresh issuance after infrastructure reconfiguration — reissue gives a clean audit boundary.
  • A first issuance that failed. A certificate tagged Issuance failed never got a certificate: the last attempt failed and nothing is retrying it. Its status stays pending because that is what it is, so reissue is how you start another attempt once the cause in Last Error is fixed. The same goes for a pending certificate whose issuance job was cancelled or cleaned up: for a pending certificate the only refusal is a fresh issuance that is still running; a retry waiting for its next attempt is moved to the front of the queue instead.
  • A manual-DNS certificate near expiry. Renew is refused for manual DNS; Re-issue is how such a certificate is refreshed (new TXT records each time).
  • Stuck renewal state. If a certificate keeps landing in renewal_failed (the safety net retries it automatically, but the root cause may make every attempt fail), reissue is the clean way to restart the lifecycle without clearing state manually.

Trigger a reissue​

  1. Open the certificate detail page.
  2. Click Re-issue (top-right actions). It is offered for ACME certificates (automatic DNS, HTTP or manual DNS) and AD CS certificates — also while they are active.
  3. Confirm. For a valid certificate the dialog says that a new key is generated and that the current certificate stays valid and is not revoked. There are no options — the key type and every other setting are inherited from the existing certificate.
  4. An issue_certificate job enqueues (msca_issue_certificate for AD CS-issued certificates, ca_place_order for a CA connector). There is no reissue_certificate job type, so do not look for one on the Jobs page. For a CA-connector certificate the confirm dialog says whether the re-issue requests a new certificate from the vendor (one unit of your quota or credit) — see CA connectors for which certificates can be re-issued.

Re-issue is refused while a renewal of the certificate is running (the API answers 409): the two would work on the same order. Wait for the renewal to finish, or cancel its job on the Jobs page, then re-issue. A re-issue requested while a manual-DNS validation or another issuance of the same certificate is already running is not started twice — the running one is reported, and two clicks in quick succession share one job. An operator whose re-issue needs approval gets every refusal at request time, with the same answer the direct path gives — nothing is filed that a reviewer could approve in vain.

Do not use the Jobs page's Retry on an old issuance job to refresh a certificate: a first-issuance job of a certificate that already holds one is refused there — Re-issue is the path (and a re-issue job stays retryable; a retried re-issue returns the certificate to its current status if it fails, not to the status it had when the job was first queued).

What happens​

  1. The worker generates a new private key of the requested type.
  2. It builds a fresh CSR with the same CN + SANs.
  3. The CA's normal issuance flow runs (DNS-01 for ACME, RPC or CES enrollment for MSCA, a vendor order — after domain validation — for a CA connector).
  4. The new cert becomes the current version on the same certificate record. The previous certificate and its private key are retained as a historical version — they are not purged, and remain downloadable per version until history retention prunes them.
  5. A cert.issued event is written to the timeline — a reissue runs the ordinary issue job, so it is recorded exactly like an issuance (there is no separate reissue event).
  6. If the certificate has auto-triggered distributions, they fire against the new cert.

While it runs the certificate shows issuing — or awaiting_validation for a manual-DNS certificate, with the new TXT records on its page — but the certificate you already have stays the one in use: it can be downloaded and distributions keep deploying it until the new one is stored. Expiry warnings and the expired notice still go out for the certificate in use while a manual-DNS re-issue waits.

Cancelling a manual-DNS re-issue. While the TXT records are pending, the DNS Challenges card offers Cancel re-issue: the pending records are discarded, the certificate returns to its previous status with its current key, the cancel is recorded on the timeline and the audit log records Reissue Cancelled. A cancel is not an error: nothing is written under Last Error and no failure banner appears. Use it when you no longer want the new certificate — or before revoking a certificate whose re-issue you started (Revoke is not offered while a re-issue waits). A manual-DNS re-issue nobody finishes is abandoned by the scheduler two hours after it started: the certificate returns to its previous status the same way, with a Certificate Issuance Failed notification; re-issue again for fresh TXT records. A running automatic re-issue (an issue job, a propagation check, a finalize or a vendor poll) is cancelled from the Jobs page instead; that cancel also restores the key, removes the TXT records it created and writes no Last Error either. If the CA had already issued the new certificate when the cancel arrived and the re-issue runs on the automatic DNS-01 flow with a separate propagation check, that certificate is kept with its new key for two hours and is not stored (its TXT records are removed); a Re-issue within that time stores it without placing a new order. After the two hours it is discarded and the certificate in use gets its own key back. A worker that stops or loses its job is not a cancel: the job runs again and continues. On the other flows (HTTP-01, RFC 2136 and the other providers that validate in one step) a certificate issued at that moment is discarded with its new key, and the next Re-issue places a new order.

If the certificate in use expires while the re-issue waits, a failed or cancelled re-issue returns it to expired, not active.

If the re-issue fails (a refused order, a failed validation, an expired challenge) nothing is replaced: the certificate returns to the status it had — an active certificate stays active with its own key — the reason is shown under Last Error and in a last re-issue failed banner, a cert.failed timeline event is written and a Certificate Issuance Failed notification is sent. Use Re-issue again once the cause is fixed.

Old cert is NOT revoked automatically

Reissue replaces the cert on our side but the CA still considers the old cert valid until it expires naturally. For key-compromise workflows, follow up with Revoke on the old cert — CertAutoPilot keeps a copy of the old cert's issuance record for this purpose in the timeline. The previous version stays in the certificate's history and remains downloadable until history retention prunes it.

Pinned clients​

If downstream clients pin the certificate's public key (HPKP, mobile apps with cert-pinning, banking integrations), a reissue breaks them. CertAutoPilot has no way to know this — there is no "pinned" annotation and no warning in the confirm step. Track pinned consumers yourself and plan the client-side update alongside the reissue.

Troubleshooting​

"Rate limit exceeded" on immediate reissue​

The CA's weekly quota for the registered domain is saturated. Wait for the counter to reset — CertAutoPilot does not display CA rate-limit counters, so check your CA's own status page — or use a staging account to avoid production consumption during testing.

Re-issue accepted but nothing changed​

Before 1.6.1 a re-issue of an active certificate completed at once without issuing anything. Since 1.6.1 the serial number and fingerprint change when the new certificate is stored; if they do not, open the job from the certificate's Jobs tab and read its log.

MSCA reissue fails with "template has recently been issued to the same subject"​

Some Windows CA templates configured with strict duplicate detection reject same-subject issuance within a window. Adjust the template's re-enrollment policy, or issue against a different template.

See also​

Video tutorial

Short, hands-on videos showing this feature working for real. English audio, English and Turkish subtitles. All video tutorials →

Re-issue a Certificate with a New Private KeyYouTube ↗