FAQ - common errors

"Monitor limit exceeded" when creating a monitor

Your workspace has reached the monitor limit for your current Uptime plan. Options:

• Delete unused monitors under Monitors > select monitor > Delete.
• Upgrade your Uptime plan: Settings > Plan.

See Settings → Usage for the workspace's live limit and pricing for the current offer.

"Feature not available on your plan"

The feature requires a capability that is not active for this workspace. The error may include the required plan. Check Settings → Plan and pricing.

Monitor is "up" but my site is actually down

Possible causes:

• Monitor URL is wrong. Verify the URL, including https:// prefix and any path.
• The monitor is checking a different endpoint. The health endpoint you are checking may return 200 even when the application is broken. Consider adding a content check (keyword match) in the monitor settings.
• Failure not yet confirmed. The monitor's confirmation settings can require more than one failure before an alert opens.
• Geographic routing. The prober is connecting from a specific region. If your DNS or CDN routes that region to a different server, you may see a different result than a local curl.

Monitor is "down" but my site is fine

This is a false positive. Common causes:

• Temporary network issue in the prober region. Multi-region verification can distinguish a regional result from a broader outage when the capability is enabled.

• Timeout too short. If your endpoint occasionally takes more than the configured timeout, it registers as a failure. Increase the timeout under monitor settings.

• IP blocking. Your server, CDN or WAF may be blocking our probers. The addresses we probe from are published at howlops.com/bot, and as JSON at howlops.com/bot/ips.json if you want your firewall to read them directly. Add them to your allowlist. Note that a block usually looks like a 403, not a timeout: if your monitor reports 403 from every region while the site loads in your browser, that is almost always this.

  You do not have to wait for a page to find out. Run test on the monitor form, when creating or editing, tells you on the spot when a target is refusing automated checks rather than being down — and names the gatekeeper (Cloudflare, Akamai, Imperva and the rest) when the response identifies one. It never stops you saving the monitor; you may be watching a site you know refuses bots. It just puts the fact in front of you while the decision is still yours to make.

• TLS certificate issue. Probers validate certificates against public certificate authorities. A self-signed certificate, a certificate signed by a private/internal CA, an expired certificate, or one with a hostname mismatch is treated as a failure. Check Monitor > SSL tab for certificate details.

"Rate limited" error (429)

You have exceeded a limit for that endpoint. Wait for the Retry-After delay before retrying. See API errors.

Heartbeat not resetting after a ping

Possible causes:

• Wrong slug. Verify the URL matches /api/v1/hb/<slug>. The slug is the credential and is case-sensitive.
• Paused heartbeat. A paused heartbeat rejects pings with HTTP 410.
• Rate limit. Each slug accepts up to 60 pings per minute. A ping may use GET, POST, or HEAD and does not require a session or API token.

"Workspace not found" after following an invitation link

• Invitation links expire after 7 days. Ask the inviter to resend.
• If you followed the link but the workspace still does not appear, you may have accepted it into a different account. Try signing out and signing back in with the email address the invitation was sent to.

Data export download link expired

Export download links expire after 48 hours. Request a new export from Settings > Account > Data export > Request export.

Status page custom domain shows a certificate error

Open the page configuration and compare the current domain state with the DNS record HowlOps asks you to publish. Verify the public CNAME with:

dig status.example.com CNAME

See Set a custom domain for the full setup guide.

API returns 404 for a resource I know exists

HowlOps returns 404 for unauthorized access to prevent tenant data leakage (IDOR protection). Verify:

• You are authenticated as the correct user and workspace.
• The resource belongs to the workspace your token is scoped to.
• The resource ID is correct (UUIDs are case-sensitive).

How do I avoid false alarms during deployments?

Create a bounded deployment maintenance window. The deployment API returns a one-time token that a cleanup step can use to end the window, and the original deadline remains the fallback. See Wrap a deployment in maintenance.

SSL certificate check is failing but my certificate looks valid

Probers validate certificates against public certificate authorities. Common causes:

• Self-signed certificate: not trusted by public CAs.

• Incomplete chain: a missing intermediate CA. Test with:

  bash

  openssl s_client -connect yoursite.com:443 -verify_return_error
  

• Hostname mismatch: the certificate CN/SAN does not match the monitored URL.

Check Monitor → SSL for the certificate details HowlOps observed.

Can I monitor internal or private endpoints?

Not directly. HowlOps checks run from external cloud locations, so private services (192.168.x.x, 10.x.x.x, or anything behind a VPN) are not reachable from the probers. Options:

• Expose a dedicated public health-check endpoint.
• Use heartbeat monitoring: have a job inside your network ping HowlOps, so reachability is proven from the inside out.

Email alerts paused: "email daily send budget exceeded"

Your workspace reached its daily non-critical email budget for that UTC day. This is an anti-abuse backstop and usually points to an alert storm.

What happens: further non-critical email alerts are paused until the next UTC day. You'll see a note on the incident timeline and an in-app notice. Critical alerts and your other channels (SMS, Slack, webhooks, …) keep working.

What to do:

• Find the noisy monitor(s) driving the volume and mute or fix the underlying flapping.
• Review routing, dependencies, and noise control so one underlying failure does not fan out unnecessarily.
• Reduce the number of email recipients on high-frequency alerts, or route them to a chat channel.
• If your normal volume genuinely needs a higher ceiling, contact support to have it raised.

See also

• API error codes reference
• Plan limits reference
• Missing notifications
• Login issues
• Prober offline