Dashboard and activity
The dashboard summarises the last 24 hours of traffic across every cluster you can see.
Alerts
Alerts is at the very top of the dashboard when something needs attention: licence seats that are expiring, in grace or expired, sentries with no seat, licensed sentries that have not checked in for 10 minutes, a past-due payment (see If a payment fails), and Website protection capacity at 80% used or full (see Website protection capacity alerts). If there is nothing to report, the Alerts section is hidden. See Licences.
Totals
The top cards show the last 24 hours:
- Passthrough HTTPS, 24h — HTTPS that is passed through unopened to your origins, including their
wwwversions: domains with Website protection off, or domains with Website protection without a usable certificate or sentry HTTPS key. Connections for hostnames you haven't added, or with no hostname, aren't counted, and a connection whose TLS handshake fails isn't counted. These aren't HTTP requests. Under the figure, from N IP addresses is how many different visitor IP addresses opened those connections. The line below it reads Connections to domains without Website protection, sent to the origin unopened. When the figure is 0 and you have domains with Website protection, it reads None. Your HTTPS traffic is inspected and counted in web requests. A 0 here doesn't mean your sites had no visitors: inspected HTTPS is counted in the web request totals instead. - Allowed web requests, 24h — HTTP requests, and inspected HTTPS requests, to all your domains that were let through without a block reason. This includes your sites' own error responses, such as a 404 for a missing page.
- Blocked web requests, 24h — HTTP requests, and inspected HTTPS requests, to all your domains that stillorigin blocked or answered with an error itself. Your sites' own error responses aren't counted as blocked.
- Open tickets — click the card to open Tickets.
Request totals count individual requests, not unique visitors. HTTPS that is passed through unopened counts as connections, not requests. Inspected HTTPS counts in the request totals, exactly like HTTP, and Passthrough HTTPS, 24h for those domains drops to about zero. See Domain traffic columns and HTTPS inspection.
Firewall blocks aren't in these totals. You'll find them on the Firewall blocks chart and under Logs. Both show a sample, so they won't match the blocked web requests total. Cluster limits count new connections per IP. See Scraper throttling.
For the same figures for a single domain, click the domain on Domains (see Domain traffic graphs).
Activity chart
Web requests, last 24 hours shows the requests your sentries can read, to all your domains, with Website protection on or off, as one stacked bar per hour:
- Mint — allowed requests
- Coral — blocked requests
A sentry can read a request when it isn't encrypted, or when it holds the domain's certificate:
- Plain HTTP on port 80, for any domain. Most of it is usually visits to
http://addresses, from people, links and crawlers, that your sites then redirect to HTTPS, plus certificate checks (for example AutoSSL or Let's Encrypt) and scanners. - HTTPS on domains with Website protection whose HTTPS is inspected (see HTTPS inspection).
HTTPS that is passed through unopened, which is most HTTPS on domains with Website protection off, can't be read, so it's counted under Passthrough HTTPS, 24h instead.
The bars cover a rolling 24-hour window and are labelled with the date and your local time. Quiet hours still get a bar, so there are always 24. Hover a bar for the exact counts. If there's nothing to show yet, the chart shows an empty state instead.
This chart doesn't include HTTPS connections that are passed through unopened, or firewall blocks.
The same request chart appears on a sentry's Health view, next to a chart of that sentry's 1-minute load average.
Traffic by country
Traffic by country shows requests to your domains, by the visitor's country. HTTPS connections to domains without Website protection go straight to the origin server and aren't counted. No IP addresses are stored or shown here. It covers every domain you've added, and only hostnames you host. Your sentries look up the country themselves.
Choose what to show (Allowed, Blocked or Block rate) and a time range (24 hours, 7 days or 30 days). Hover a country for its counts. Top countries under the map lists the same numbers.
- Allowed and Blocked shade a country from its first request.
- Block rate is shown only for countries with at least 20 requests in the selected range.
- Requests with no known country, such as from private networks, and countries with very little traffic (grouped as Other countries) are listed under the table. They're never coloured on the map.
Firewall blocks
Firewall blocks shows a sample of firewall blocks across your clusters, stacked by reason: rate limit, temporary ban, deny list, connection limit, crawler limit, origin limit and identified traffic block. It isn't every dropped connection, and it isn't the same as the blocked HTTP requests total in the cards above.
Choose a time range on the chart. The last 24 hours are shown by the hour, the same rolling window as the web requests chart. Longer ranges are shown by the day. The 7 and 30 day views come from a daily summary that updates about once an hour. Open Logs for the individual events.
Identified traffic
Below the activity chart, Identified traffic, last 24 hours groups traffic by type. A horizontal stacked bar shows allowed versus blocked for each group, and the table below lists every Source with Allowed and Blocked counts. Allowed and blocked counts include both HTTP requests and HTTPS connections.
The groups are:
- Search engines — verified search engines, such as Googlebot and Bingbot. This group starts with crawler limits on.
- AI crawlers — verified AI crawlers, such as GPTBot and Claude. This group also starts with crawler limits on.
- Monitoring & uptime — uptime and monitoring services we can identify from their official IP lists.
- Site management — site management tools, such as WordPress management services.
- Payment providers — payment providers' notification and webhook servers, such as Stripe and PayPal.
- Security platforms — security services that publish their own address lists.
- Internet scanners — services that scan the internet, such as Censys. They are labelled, never trusted, and you can block them.
- Origin servers — traffic from your origin servers. This group was previously called Origin / host monitors.
- Unverified — all other traffic to hostnames you host.
- Unknown host — HTTP requests for hostnames you don't host, and HTTPS connections with no hostname, a hostname you don't host, or just the IP address. HTTP requests for hostnames you don't host are answered with 403 and counted as blocked.
- Firewall blocks — in their own rows: Rate limit, Temporary ban, Deny list, Connection limit, Crawler limit, Origin limit, Identified traffic block and Country block.
Crawler limits are a per-group setting. Search engines and AI crawlers start with them on; the other groups start with them off and use the ordinary limits. Only stillorigin staff can change which groups get crawler limits.
You can block a group or a single service on Block identified traffic. Being named never lets traffic skip your limits, the WAF or your deny list, and identified traffic is never added to the naughty list automatically.
The same breakdown is available:
- on the dashboard, across all clusters;
- on Clusters, when you expand a cluster;
- on a sentry's Health view, for that sentry only;
- on a domain's page, as Identified traffic on this domain. That view only shows traffic on that domain's hostname. Firewall blocks with no hostname appear only in the cluster totals.
See Verified crawlers, Block identified traffic and Scraper throttling.
If a service your customers rely on is missing, use Missing a service? Request it to open Request a service.
Request a service
Request a service is for signed-in hosts. Tell us about a service that's missing from Identified traffic. We check every request against the requirements below, and when a service qualifies, we add it for every stillorigin customer.
The form asks for:
- Service name
- Service website
- Official IP list URL (if you know it) — optional. If you're not sure, leave it blank and we'll look for it
- Type of service
- Why you'd like it added — optional
You can send up to 5 requests a day, and have up to 10 waiting for review. Each request has a status: Received, Under review, Added or Not added, and may include a note from us. Requests appear in your audit log.
What a service needs before we can add it
- The service publishes its IP addresses itself, on its own website or at an address its own documentation links to. A list copied by a third party, a forum post or a blog isn't enough.
- The list is at a fixed address we can check automatically, without signing in.
- The list is in a format a computer can read, such as a plain list of IP addresses and ranges, or JSON.
- The addresses are used only by that service. A list that covers a whole cloud provider, CDN or shared network can't qualify, because a match wouldn't prove the request came from that service.
- The service keeps the list up to date. We check each list about once a day.
- The addresses are the ones the service uses to connect to websites and servers.
- If the service connects over IPv6, the list includes its IPv6 addresses too, or the service says it uses IPv4 only.
Adding a service names its traffic in Identified traffic. It doesn't let that traffic skip your limits, the WAF or your deny list.
Why some services aren't included
- There's no official list — some services don't publish their IP addresses, or say their addresses can change at any time. Let's Encrypt is one example. Shopify, Slack and Zapier ask you to check a signature instead. We won't guess or copy a list from somewhere else.
- The addresses are shared — Cloudflare health checks come from Cloudflare's own network, and those same addresses carry traffic for millions of other websites and Cloudflare products. The same goes for whole cloud providers such as AWS, and for Datadog's full list. A match can't prove the request was a health check from that service, so naming it would be misleading.
- The list covers more than the service — some vendors publish one list for all their systems, such as update servers, licence checks and support tools, so we can't label all of that traffic as one service.
- Internet scanners are only labelled — services such as Censys scan the whole internet. We can name their traffic so you can see it, and you can block it if you want to, but we never treat it as trusted.
Block identified traffic
On Settings › Identified traffic, the page headed Block identified traffic lets you block services we identify by their official IP addresses, such as AI crawlers, across your whole network or on one cluster.
Only an account owner can change these blocks. Other users on your account can see the page but can't change it.
Use the All clusters tab for blocks that apply to every cluster in your account, now and in future. A cluster tab's blocks apply to that cluster only, on top of the blocks for all clusters. Blocks for all clusters are shown ticked and can only be changed on the All clusters tab. From a cluster's Edit page, Identified traffic blocks for this cluster opens that cluster's tab.
You can Block the whole group or block a single service. Blocked connections are dropped before they reach your sites, and show as blocked in Identified traffic and Logs. Your origin servers, sentries and cluster VIP are never blocked this way.
Click Save blocks to apply your changes. Sentries pick them up within a minute. Blocking needs the latest sentry release. Sentries that haven't updated yet keep letting this traffic through.
If you're about to block a search engine, the page warns you: blocked search engines can't crawl your sites, so pages can drop out of search results. Click Block anyway if you're sure.
Under Default blocks for your clients' domains on the same page, you can set default blocks for your clients' domains with Website protection. Visitors from a blocked service get the stillorigin block page, and a domain owner can allow that service on their own domain after reading a warning. Defaults don't apply to domains without Website protection. Network blocks on this page are different: they drop the connection for every domain and stay locked for owners. See Default blocks for owners.
Fake crawlers
Block fake crawlers blocks requests that say they're a known crawler, such as Googlebot, Bingbot or GPTBot, but don't come from that crawler's official IP addresses. This works on domains with Website protection, and on plain HTTP for other domains. HTTPS without Website protection isn't checked, because the sentry can't see the request. It always blocks, even while Website protection rules are in log mode. See fake-crawler.
You can turn this on for all clusters or for one cluster, the same way as the other blocks.
Domain owners
Owner access comes with every domain with Website protection. You can allow domain owners to block identified services on their own Website protection domain with the permission Block identified services, such as AI crawlers, on their domain (domains with Website protection only). Like the other owner permissions, it is off by default. See Domain owner access. Domain owners can also allow a service you block by default, without that permission. See Default blocks for owners.
Country blocking
On Settings › Country blocking, you can block visitors by the country their IP address is in, or ask them to complete a Captcha Protection check on domains with Website protection. Network-wide blocks cover every domain, including domains without Website protection: HTTPS is stopped at the sentry's firewall, and plain HTTP on port 80 is stopped by the sentry's web server, which always lets certificate checks through. New countries start in log mode. See Country blocking.
Captcha Protection
On Settings › Captcha Protection, you can ask visitors to your clients' domains with Website protection to complete a quick, automatic browser check before the page opens, and turn on Under Attack in an emergency. See Captcha Protection.
Licence alerts
The alerts list warns when a sentry, origin, or Website protection-domain seat is expiring soon (within 14 days), in its grace period, or expired, and when a sentry has no seat assigned. See Licences. Open tickets are not listed as alerts; use the Open tickets card.