Troubleshooting by symptom

Capture evidence first#

Record the affected URL or operation, UTC timeframe, expected and actual result, screenshot when useful, request/trace ID, workspace/project/monitor context, and current release/environment. Never send secrets, tokens, private heartbeat URLs, or full customer data.

Monitor has no data or stays Pending#

• Confirm the monitor is Active and the project is not archived.
• Wait at least one configured interval; TLS, Domain, and Sitemap use fixed lower cadences.
• Open the monitor detail and check whether it is waiting for a fresh result after a settings change.
• Confirm the URL and monitor type are valid.
• Review Watchfox status and contact support if multiple unrelated monitors stop updating.

Monitor shows Down#

• Review the latest non-OK status, latency, final URL, redirects, and error code.
• Test the target from another independent connection.
• Check DNS, origin availability, CDN/firewall rules, and whether automated requests are blocked.
• Confirm Expected HTTP status is not stricter than intended.
• Avoid changing the monitor until you understand whether the target or the rule is wrong.

Redirects or redirect limit#

A normal redirect can be healthy when 200–399 is accepted. A redirect loop, missing Location header, or too many redirects is a target/configuration problem. Check the final canonical URL and remove unnecessary redirect chains.

Timeout#

A timeout means Watchfox did not receive a response before the configured limit. Compare the normal target latency, origin load, firewall behavior, and timeout value. Increasing the timeout can hide a real performance problem, so change it only when the slower response is acceptable.

Keyword mismatch#

• Confirm the exact rule and Should contain/Should NOT contain logic.
• Inspect final URL, cache evidence, response age, body size, and available headers.
• Check whether the text is rendered only by JavaScript or requires login/cookies.
• Purge stale CDN or WordPress cache when Watchfox is correctly receiving an old response.
• Read Keyword monitoring.

Keyword response too large#

too_large means the response exceeded the 256 KiB inspection limit and the keyword rule was not evaluated. Use a smaller endpoint, reduce the response, or switch to Uptime when content inspection is not required. See Response too large.

TLS warning or Unknown#

Check the intended HTTPS hostname, certificate chain, expiry evidence, and whether the source returned a definitive result. Unknown/source-limited is not the same as confirmed expiry or outage.

Domain expiry Unknown or source-limited#

Registry and RDAP coverage varies. Verify the domain at the authoritative registrar/registry and keep renewal ownership current. Do not treat Unknown as a confirmed expired domain.

Sitemap broken or redirected URLs#

Open representative findings, confirm whether redirects are intentional, and update the sitemap to canonical URLs where appropriate. Remember that Watchfox validates the supported sitemap scope; it is not a full deep crawler.

Heartbeat says Missing#

Confirm the job schedule, last successful run, integration URL, expected interval, and whether the ping happens at successful completion. A missing ping can mean the job failed, never ran, timed out before the ping, or lost access to Watchfox.

Test alert or email did not arrive#

1. Check Watchfox channel state and recent test/delivery result.
2. Confirm the exact environment, recipient, and UTC timeframe.
3. Check provider activity and suppressions only after confirming Watchfox attempted delivery.
4. Check spam, inbound filtering, mailbox validity, and recipient spelling.
5. Distinguish fair-use quota suppression from provider failure.
6. Reset delivery state only after the underlying cause is understood.

Report does not contain expected data#

Confirm project, period, included monitor types, enabled state, generated-at/data-as-of timestamps, and retention window. Uptime metrics include enabled Uptime and Keyword monitors; old raw check rows may be absent even when rollups, daily summaries, and incidents support the report.

Contact support#

Use the in-app Contact support action or email [email protected]. Include the minimum evidence needed to reproduce the issue and avoid secrets or private customer content.