Australian cloud for business software — built and run in Sydney since 2002

NetLeverage
NetLeverage

Let's Encrypt "too many certificates ... already issued": which rate limit, and how long to wait

Technical article
Troubleshooting

About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.

Let's Encrypt returns "too many certificates (N) already issued" when you have used up one of its certificate rate limits. The message names the limit and ends with a time to retry after. Two limits produce this wording: certificates per exact set of names, and certificates per registered domain. The numbers below come from the Rate Limits page as last updated on August 5, 2026. Limits change, so check the current page before relying on them.

What the error means

Let's Encrypt calculates limits "per request, using a token bucket algorithm". You can make requests "in bursts—up to the full limit—or space out your requests". Capacity refills over time. The page is direct about one thing: "If you've hit a rate limit, we don't have a way to temporarily reset it." It adds that "Revoking certificates does not reset rate limits".

The error text is built in Boulder, the software behind Let's Encrypt. In the main branch source, the two certificate messages are:

  • too many certificates (%d) already issued for this exact set of identifiers in the last %s, retry after %s
  • too many certificates (%d) already issued for %q in the last %s, retry after %s

The first is the exact-set limit. The second names a registered domain (or another network unit). The retry time is in UTC. The source adds a random 0 to 3 percent to the wait to avoid every client retrying at once, so do not expect the stamp to match your arithmetic to the second. Messages follow this format, and where more than one limit is exceeded, the page says the error is always for "the limit that resets furthest in the future". Responses also carry a Retry-After header.

The text above is from one undated branch of the source, so match on "too many certificates" and read which phrase follows.

The limits that matter here, from the August 5, 2026 page:

Limit Value on that page Refill rate on that page Check on the arithmetic
New Certificates per Exact Set of Identifiers 5 per 7 days 1 certificate every 34 hours 7 days = 168 hours; 168 ÷ 5 = 33.6 hours, shown as 34
New Certificates per Registered Domain 50 per 7 days 1 certificate every 202 minutes 7 days = 10,080 minutes; 10,080 ÷ 50 = 201.6, shown as 202
New Orders per Account 300 per 3 hours 1 order every 36 seconds 3 hours = 10,800 seconds; 10,800 ÷ 300 = 36
Authorization Failures per Identifier per Account 5 per hour 1 per identifier every 12 minutes 60 minutes ÷ 5 = 12

The page says the exact-set limit "is a global limit, and all new order requests, regardless of which account submits them, count towards this limit", and that there are no overrides for it. An override for the registered-domain limit can be requested, but "It takes a few weeks to process requests", so it is not a way to reset a limit quickly.

Renewals are treated specially. The page says a renewal coordinated through ACME Renewal Info (ARI) is "exempt from all rate limits". A non-ARI order with exactly the same set of names is exempt from the orders and registered-domain limits, but "would still be subject to" the exact-set and authorization-failure limits.

Likely causes

Cause What you see How to confirm
The same names are requested again and again The message says "for this exact set of identifiers" The Certbot logs and your deployment history show repeated requests for the same names
Client state is lost on each deploy A new certificate is requested after every redeploy or container start /etc/letsencrypt is missing or empty after a deploy
Renewal forced on a schedule A cron entry or timer runs with --force-renewal The scheduled command line
Many certificates under one registered domain The message names a domain in quotes The Certificate Transparency search shows many recent certificates for the domain
Validation failures pile up A different message about failed authorizations The Rate Limits page entries for authorization failures

Prove the cause before changing anything

All of these are read-only.

1. Read the whole message. Note the number in brackets, the phrase after "for", the period (the source prints it with %s, so a seven-day limit reads as a count of hours, as in the example below) and the "retry after" time.

2. List what Certbot has.

certbot certificates

The Certbot guide describes the output: each certificate's name, domains, expiry and paths. A certificate that should exist but is missing points at lost state.

3. Read the Certbot log. The Certbot guide says it "stores status logs in /var/log/letsencrypt" by default. Look for repeated issuance requests for the same names.

4. Find what runs the client. The guide says to look for the certbot renew command "in either your system's crontab (typically /etc/crontab or /etc/cron.*/*) or systemd timers (systemctl list-timers)". Also check deployment scripts, container start-up scripts and any control panel that issues certificates.

5. Count outside your own records. The Rate Limits page says you can list certificates issued for your registered domain "by searching crt.sh or Censys, which use the public Certificate Transparency logs". The count there is the one Let's Encrypt applies, across all accounts and all servers.

Fix

Risk: the steps below that request certificates change what Let's Encrypt counts against you, and some can trigger web server reloads. Do not run an issuance command until you know why the limit was reached.

Wait for the retry time

If the exact-set limit was reached, the same set of names can be requested again after the retry time. Until then, the existing certificate, if it is still valid, keeps working. Check its expiry with certbot certificates and compare it with the retry time.

Stop the loop

First name the object. Run certbot certificates and note the expiry date of the certificate, then systemctl list-timers and read the crontab to find the timer, cron entry or script that requests certificates. Copy a script you are going to edit to a dated name in /root (for example cp -p deploy.sh /root/deploy.sh.2026-11-20), and write down the exact name of any timer you will stop.

Risk: stopping the Certbot timer stops renewals, and a certificate that is not renewed expires on the date you noted. Stop only the entry that forces issuance, not the normal renewal timer, and restart anything you stop before that date. To back out, restore the saved script under its saved name, or start the timer again with systemctl start and its exact name.

Then find and stop whatever keeps requesting certificates: a deployment step that deletes the Certbot configuration, a container that issues on start, or a timer with --force-renewal. The Certbot guide says --force-renewal is "not appropriate to run daily because each certificate will be renewed every day, which will quickly run into the certificate authority rate limit". The Integration Guide describes the container case: "if you have to destroy and re-create all of your instances at once, you may wind up in a situation where none of your instances are able to get a certificate". Keep /etc/letsencrypt on durable storage and restore it before deploying.

Test against staging

The Staging Environment page says to use --test-cert or --dry-run with Certbot and warns that "ACME accounts are scoped to each environment". Staging uses the same limits "but with different values"; its exact-set limit is "30000 per week". Before a dry run, print the current state and run your web server's configuration check (for example nginx -t or apachectl configtest), so you can compare afterwards:

certbot certificates

Risk: Certbot notes that a dry run "may trigger webserver reloads to temporarily modify & roll back configuration files". Run it in a quiet period. There is nothing to restore afterwards; run certbot certificates and your web server's configuration check again and compare them with the output you printed before.

For testing, use:

certbot renew --dry-run

The Certbot guide says --dry-run uses the staging server, and warns that with a custom server "it may cause real certificates to be issued and discarded". Do not combine --dry-run with --server unless you know the target.

Change the set of names (only if you can accept it)

The Rate Limits page offers a workaround: "you can change the set of identifiers by adding blog.example.com to request additional certificates". It adds that these new orders "would not be considered renewals" and so count against the orders and registered-domain limits. First record what you have and save the renewal file:

certbot certificates > /root/certbot-certificates.2026-11-20.txt
cp -p /etc/letsencrypt/renewal/example.com.conf /root/example.com.conf.2026-11-20

Risk: this issues a new certificate with a new set of names and counts against your limits. It cannot be undone cleanly: returning to the old set is another new order, and under the exact-set limit that is the set you just exhausted. Your web server also needs a reload before it serves the new file; see "Certbot renewed but the site still shows the old certificate: reload, deploy hooks and the right file". Do this only when the extra name is acceptable.

The Certbot guide's "Changing a Certificate's Domains" section says the --cert-name flag "can also be used to modify the domains a certificate contains, by specifying new domains using the -d or --domains flag", and its example is certbot certonly --cert-name example.com -d example.com. Its section on renewing with certonly adds that "All of the domains covered by the certificate must be specified", and shows -d repeated once per name (certbot certonly -n -d example.com -d www.example.com). So give the full list you want, one -d per name. For example, to add a name to a certificate that already covers example.com and www.example.com:

certbot certonly --cert-name example.com -d example.com -d www.example.com -d blog.example.com

Example (illustrative): web01 runs a deployment script that deletes /etc/letsencrypt and then runs certbot certonly for app.example.com and www.example.com on every release. On the sixth release in a week the log shows too many certificates (5) already issued for this exact set of identifiers in the last 168h0m0s, retry after 2026-11-27 04:12:00 UTC. The current certificate is still valid for 19 days. The administrator removes the deletion step from the script, restores /etc/letsencrypt from backup, and tests the script's issuance step with --dry-run. On the retry date, the next issuance succeeds, and later renewals follow the normal timer.

Confirm the fix

  1. After the retry time, certbot renew --dry-run completes without a rate-limit message. The dry run uses staging, so it proves the client and the web path, not that production capacity has refilled.
  2. certbot certificates shows the certificate you expect with an expiry date in the future.
  3. The next scheduled renewal does not request a certificate for a set of names that already has a valid one.
  4. The served certificate matches the new file: "Certbot renewed but the site still shows the old certificate: reload, deploy hooks and the right file" shows the check.

What to watch afterwards

  • Alert on renewal failures from the timer, not on the day the certificate expires.
  • Never delete the Certbot configuration directory as part of a deploy. Back it up like any other state.
  • Count certificates per registered domain if you issue one per customer or per site, and compare with 50 per 7 days.
  • Re-read the current Rate Limits page before a bulk change, a migration or a mass re-issue.

Risks and limits

  • These commands have not been run or lab-tested.
  • The limits and refill rates are from the page dated August 5, 2026. Check the current page.
  • The error text comes from the main branch of the Boulder source, which carries no date. The service's wording can change between releases.
  • Other limits exist: new registrations per IP address, per IPv6 range, consecutive authorization failures, and per-endpoint request limits. Only the certificate limits are covered in detail.
  • The ARI exemption applies only to clients that support ARI. Let's Encrypt's Rate Limits page does not list which clients support ARI.

Sources

  • Rate Limits — Let's Encrypt, Last updated: August 5, 2026, accessed 20 November 2026
  • Staging Environment — Let's Encrypt, Last updated: April 10, 2026, accessed 20 November 2026
  • Integration Guide — Let's Encrypt, Last updated: June 23, 2025, accessed 20 November 2026
  • Boulder source: ratelimits/limiter.go — Let's Encrypt (ISRG), main branch, no date shown, accessed 20 November 2026
  • Certbot User Guide — Electronic Frontier Foundation, stable documentation, no date or version shown, accessed 20 November 2026