How SEOComet works
Checks run on a schedule and report issues. Most issues open an incident, which alerts you when it opens, gets worse and is resolved.
Updated
SEOComet fetches your site’s files and pages on a schedule and compares what it finds with what it found before. When something breaks, it opens an incident and tells you through your alert channels until the problem is gone.
Checks
A check is one file or page that SEOComet fetches on a schedule, and the rules it applies there. There are 6 kinds:
- robots.txt
- Sitemap
- ads.txt
- app-ads.txt
- Uptime and pages
- SSL certificate
A site usually has several checks: one robots.txt, one per sitemap, ads.txt and app-ads.txt if the site has them, an uptime check on the home page and an SSL check. You can add more sitemaps and pages yourself. A sitemap with Google News tags shows as News sitemap, and an uptime check shows as Homepage on the home page and Page anywhere else.
How often each check runs depends on your plan. See How often checks run.
Each run ends one of three ways:
- OK: nothing to report.
- Issues found: the check found one or more issues.
- Couldn’t check: SEOComet couldn’t get a usable answer from the address, such as a timeout or a server error. See A check says “Couldn’t check”.
Issues
An issue is one thing a check found, such as “Search engines blocked” or “Sitemap looks stale”. Each has a severity: critical, warning or info. A few are suggestions, which are worth tidying but break nothing. Severity and suggestions explains each level.
A check’s status comes from the worst issue in its latest run. A check whose only issues are info or suggestions counts as healthy.
Incidents
An issue that isn’t a suggestion opens an incident. An incident tracks one problem on one check from the run that found it until a run no longer finds it.
- A check has at most one open incident per kind of issue. When the next run finds the same issue, it’s added to the open incident, not reported as new.
- A file change found together with a warning or critical issue is attached to that issue’s incident instead of opening its own.
- If the issue gets more serious, such as a warning that becomes critical, the incident gets worse with it. It never drops back to a lower severity while it’s open.
- The incident closes on its own when a run no longer finds the issue. A run that couldn’t fetch the address doesn’t close anything, since it couldn’t see.
- Someone with Can edit can also mark it resolved, or mute it so it stays open without alerts.
See Incidents for muting, reminders and what each incident page shows.
Alerts
An alert is a message about an incident, sent to your channels: email on every plan, and Slack or a signed webhook on Pro and Business. Alerts go out when an incident opens, gets worse and is resolved. If you turn reminders on, a “Still open” reminder also goes out every 4 hours while it stays open. While a site’s homepage is down, its other incidents don’t send alerts, so an outage sends you one alert rather than one for every page and file. See Was the site down?.
Your account’s alert level decides which incidents send alerts: Critical only, Warnings and critical or Everything. A site can have its own level, or have its alerts turned off. A muted incident sends nothing until it’s resolved. See Alert level and per-site alerts and Alert channels.
Needs attention
Needs attention on the Overview lists the checks whose latest run is critical, warning or couldn’t check, worst first. Only running checks on running sites are listed, so paused ones don’t show. Each row says how long the check has been failing. The Overview shows the first few; the full list is on the Checks page.
Sites at risk
Sites at risk on the Overview lists the sites with at least one check that’s critical, warning or couldn’t check, worst first. Each shows how many of its checks are in each state, and how many incidents it has open. A site whose checks are all healthy, paused or not checked yet isn’t listed.
Health
A site’s health is how its checks are doing. The Sites list shows how many checks are critical, warning or couldn’t check, or Healthy when none are. A site still being set up shows Setting up, then Not checked yet before the first results, and a paused site shows Paused. Sites sort by their worst check.
Versions and compare
SEOComet keeps a copy of robots.txt, ads.txt and app-ads.txt from each check. Each different copy is a version, listed on the check’s Versions tab with when it was first seen and how many checks saw it. Compare versions shows the lines that changed between any two of them, and a change issue has a View diff button that opens the same comparison.
Versions last as long as your check history. The latest version is always kept, so the next change can be compared with it.
Check history
Every run of a check and what it found is on the check’s History tab. How far back it goes depends on your plan: 7 days on Free, 90 days on Pro and 365 days on Business.
After you move to a plan with a shorter history, the longer one is kept for 30 days, so a payment problem that gets fixed doesn’t cost you months of history.
Activity log
The activity log records who changed what in the account: sites, checks, people, alert settings, channels and incidents, including checks someone ran by hand. The owner and people with Can edit on all sites can see it. A site’s own log is its Activity page, and the whole account’s opens from Team.
It’s kept for 30 days on Free, 90 days on Pro and 365 days on Business.