Your own visits are traffic too, and so is a monitoring service checking that the site is up. Exclusions stop that being recorded.
Dropped, not hidden
An exclusion is applied at ingestion. The hit is discarded before anything is written, so it is not in your reports, not in your history, and not counted against your credits.
That also means it cannot be recovered. A rule that is too broad quietly deletes real visits, and no report will show you the gap, so add rules narrowly and check the numbers afterwards.
Where the rules live
Each site keeps its own. There is no platform-wide list, deliberately: an office address is a local fact, and a rule set centrally that removed your real traffic would be impossible to diagnose from your side.
You can exclude by:
- IP address: a single address or a CIDR range.
- Path: the part of the URL after the domain.
- Hostname: useful when a staging host serves the same pages.
- Country.
Writing a list
One entry per line. Blank lines are ignored, and a line beginning with # is a comment that never matches anything:
# The Tallinn office
90.191.0.0/16
# Uptime checks
185.199.108.0/22
Comments are worth the trouble. A list of bare addresses is a list nobody dares delete from a year later.
Paths and hostnames match with * as a wildcard. There are no regular expressions anywhere. Patterns are typed by customers and run on the ingestion path, and a pattern that could backtrack would slow every site on the server.
Excluding is not the same as filtering bots
Skomi does both, and the difference matters:
| Exclusion | Bot detection | |
|---|---|---|
| The hit is | dropped | dropped, unless you keep bot traffic |
| In your reports | absent | absent either way |
| Reversible | no | only while you are keeping it |
Track bots is the switch that decides. It is off by default, because a stored crawler hit counts against your credits like any other and no report shows it, so a bot verdict discards the hit just as an exclusion would. Turn it on and the visit is stored and labelled instead: that is what fills the Bots page, and it is what makes a mistake in the bot list cost you nothing permanent.
So with Track bots on, prefer bot detection for anything automated and recognisable, and keep exclusions for traffic you know is yours. With it off, give a bot rule the same care as an exclusion, because both throw the traffic away for good.
A spike you did not expect
A site ID is public, because it is in the snippet on your own pages, so anyone who finds it can send hits to it. Skomi watches for unusual volume and offers to exclude the source in one step. See traffic alerts for what it looks for and what it does not.
Related: installing the snippet, how long Skomi keeps your data.