Domains and clusters
Clusters
A cluster is a group of one or more sentries that share the same settings, domains and deny list. All of its sentries handle traffic; your DNS points at each sentry's public IP address (see Point DNS at your sentries). You add and manage clusters on Clusters.
Each cluster row shows its limits, how many sentries are online and how many domains it serves, whether its WAF is on or off, and its next update window. On the row you can Edit the cluster, or Delete it (see Deleting a cluster below). Force update all at the top of the page updates every sentry (see Sentry updates).
Click a cluster to expand it and see:
- an Identified traffic breakdown for the last 24 hours;
- Manage sentries and Add sentry;
- its sentries (status only), with Status (Online when seen in the last 10 minutes), Priority, Interface, licence Expires, Version, Update status, Load, Capacity (see Sentry capacity) and Last seen;
- live readings for the cluster.
To change the cluster's settings, or to use Force, Health, New key, Edit or Delete on a sentry, click Edit or Manage sentries.
Live readings
When you expand a cluster, and on a sentry's own page, tiles show live readings from each sentry's latest check-in, next to the cluster's limits:
- New connections — new connections per second from the busiest IP, against the per-IP limit, with the total and the burst.
- Crawlers — new connections per second from the busiest verified crawler IP, against the crawler limit.
- Open connections — connections open right now, with the per-IP limit.
- HTTP requests — the live HTTP request rate: requests per second from the busiest IP, against the Website protection request limit, with the total and the burst.
Readings are averaged over the last minute. Per-IP values show the busiest IP measured against each per-IP limit. For IPv6, that's the busiest /64. In the cluster view, per-IP values come from the busiest sentry, because each sentry applies the limits on its own. Totals are added up across all sentries.
Until a sentry has checked in, the tiles show the limits only. If a sentry's live figures are out of date, they're left out and the tile says so.
Adding a cluster
Use Add cluster at the top of Clusters. On the page that opens, enter a Name (letters, numbers, dots, underscores or hyphens) and, only if support has given you one, a VIP; otherwise leave it empty. After you add it, you go to the cluster's edit page. Then add a sentry to the new cluster.
You can't add a cluster while another of your clusters has no sentries. Add cluster is turned off, and the portal names that cluster: add a sentry to it, or delete it, first. Adding a cluster doesn't use a licence; each sentry you add to it does.
A cluster name must be unique in your account. Some names used by older clusters can't be reused.
Editing a cluster
Click Edit on a cluster (or Manage sentries when it is expanded) to open its page. The Settings section is where you change its name, its limits, its SSH allow-list, its extra allowed ports and whether to use the WAF:
| Setting | Default | What it does |
|---|---|---|
| New connections per second per IP | 10 | How many new connections one visitor IP may open each second |
| Burst (new connections per IP) | 50 | Extra new connections allowed at once, for example when a page first loads |
| Open connections per IP | 50 | Most connections one IP may hold open at the same time to ports 80 and 443 |
| Crawler new connections per second per IP | 2 | The same limit for verified crawler ranges |
| Crawler burst (new connections per IP) | 10 | Extra new connections a verified crawler may open at once |
| Website protection requests per second per IP | 30 | How many HTTP requests one visitor IP may make each second on domains with Website protection |
| Website protection burst (requests per IP) | 300 | Extra requests allowed at once on domains with Website protection, for example when a page loads all its images |
| Temp ban (minutes) | 10 | How long a persistent offender stays temporarily blocked |
Defaults apply to new clusters. An existing cluster keeps its current values until you change them. When a cluster's crawler limit isn't the recommended 2 per second with a burst of 10, for example on an older cluster set to 100 per second with a burst of 200, the page says so under the crawler fields. Reset to recommended fills in 2 and 10; click Save to apply them. On a cluster created before the Website protection settings existed, they start with the same values as its new-connection limits.
For IPv6 visitors, connection limits and temporary bans apply to the whole /64 network, because one IPv6 user normally has a whole /64.
VIP
A VIP is a floating IP address that moves between the cluster's sentries if one fails. Leave it empty unless your provider supports a floating IP and support has helped you set it up. With no VIP, the sentries don't run keepalived. With a VIP, the sentries talk to each other directly (unicast) once each has reported its address, so clusters on the same network don't interfere.
SSH allow-list
Enter the IPv4 or IPv6 addresses or ranges you manage the sentries from, one per line (up to 100). Add my IP puts the public address you're using to browse the portal onto the list, unless it is already covered. Private or local addresses can't be added this way: enter the public address you manage the sentries from by hand. The address is only added to the form; click Save to apply it.
When the list has at least one entry:
- only those addresses can reach SSH on the cluster's sentries (port 22 and any other port SSH listens on);
- every other inbound port is closed, except web traffic on 80 and 443 and your extra allowed ports;
- SSH sessions that are already open stay connected.
When the list is empty, SSH and other ports stay open to everyone (except your extra allowed ports) and the cluster shows SSH open to all. Check that the address you connect from is in the list before you save. If you lock yourself out, empty the list and save (the sentries open SSH again within about a minute), or sign in through your provider's console.
Extra allowed ports
Use this for other services you run on the sentries, such as a monitoring agent. Enter one TCP port or range per line, then the IPv4 or IPv6 addresses or ranges that may connect to it, then an optional note after #. For example, to let your Nagios server reach NRPE on port 5666:
5666 203.0.113.10, 2001:db8:5::10 # NRPE
- Only the addresses you list can connect to that port. A port with no addresses after it is closed to everyone. This is the opposite of the SSH allow-list, where an empty list leaves SSH open.
- This works whether or not the cluster has an SSH allow-list.
- A range can cover up to 100 ports (for example
9100-9101). You can add up to 20 lines, with up to 100 addresses or ranges each. - Ports 80 and 443 (web traffic), 81 (used by the sentry itself) and 22 can't be added. For SSH, use the SSH allow-list.
- The service itself must be installed and listening on the sentry. This setting only opens the firewall for it.
WAF
The WAF is off until you turn it on for the cluster. Each cluster row on Clusters shows whether its WAF is on or off, so you can check at a glance. To turn it on, tick Use the WAF (OWASP Core Rule Set) on domains with Website protection from the cluster's Edit page. You can still turn it off for a single domain on Edit domain, under WAF.
See Scraper throttling for what each value does. Changes reach every sentry in the cluster within about a minute. Renaming a cluster does not affect its sentries.
Sentries
The same page lists the cluster's sentries, with Add sentry at the top (see Installing a sentry). Sentry row buttons:
- Force — install the latest version now (see Sentry updates);
- Health — the sentry's load and HTTP requests over the last 24 hours, and its identified traffic;
- New key — issue a new device key for the sentry, shown once. The old key stops working at once (see Sentry device keys);
- Edit — opens a page to change its priority or interface (leave these at their defaults unless support asks you to change them);
- Delete — only for a sentry that has never checked in. It returns the sentry licence to your pool.
When you add a sentry, you land on its page, which shows its device key and install command. They stay there while you move around the portal in the same browser tab, and are gone once you refresh, sign out, close the tab or open a new tab. If you lose them, use New key (see Installing a sentry).
Sentry capacity
The Capacity column on a sentry's row shows how full the sentry is right now, as a percentage. It's based on the sentry's open web connections or its connection table, whichever is fuller. The figure is highlighted from 70%. At 80% it's shown as an alert and you get a capacity alert email. The alert clears once usage drops back below 70%. Hover the figure for the numbers and when they were reported.
This is the sentry's current capacity and load. It isn't the HTTPS connection counts on Domains.
Sentry settings
Each time your settings change, every sentry in the cluster applies them. The Settings section on a sentry's page shows how that went:
| Status | What it means |
|---|---|
| Applied | The sentry is using its latest settings. |
| Not applied | The sentry couldn't apply its latest settings. It's still protecting your sites with its previous settings. |
| Not protecting | The sentry couldn't apply its latest settings and has never applied any, so it isn't protecting any sites yet. |
A sentry that couldn't apply its settings tries again every 10 minutes and whenever the settings change. It's also flagged with a Settings not applied or Not protecting badge on Clusters and an alert on the dashboard. If the status doesn't clear, contact stillorigin support and include the Error reported by the sentry shown on its page.
Deleting a cluster
A cluster with no sentries can be deleted after you confirm. The confirmation lists the domains and deny-list entries that will be removed with it; Website protection licences on those domains return to your pool, and origin servers are unassigned from the cluster, not deleted. Past logs are kept.
Origins
Origins lists the web servers your domains are sent to. For each origin you set a Name, Upstream IP, HTTP port (the port the sentry sends plain HTTP and Website protection traffic to, usually 80), HTTPS port (the port HTTPS is passed through to for domains with Website protection off, usually 443) and Cluster. The origin server's type is set by the origin agent: it detects cPanel, Plesk and DirectAdmin, and everything else is Generic. For Enhance, tick This origin server uses Enhance before you copy the install command. You can't change the type in the portal. New origin servers start as Generic until the agent reports.
The list also counts domains for each origin server:
- On server — the domains on the server itself, as its origin agent last reported them (a domain and its
www.count once). It shows - until an origin agent is installed and has reported. - Network protection — your domains on this origin with Website protection off.
- Website protection — your domains on this origin with Website protection on.
If On server is higher than Network protection and Website protection together, some domains on the server may not be protected yet: open the origin to see them under Not protected yet.
The origin page shows the origin agent's status (Connected, Not connected or Not installed), Last seen {when} or Not seen yet, and its version. When an origin agent is installed, Not protected yet lists domains it found on this origin server that you haven't added. You can search the list, see whether a certificate was Found, and add domains to a cluster as Network protection or Website protection (starts in log mode), up to 500 at a time. They aren't protected until you add them. See Not protected yet.
Each sentry includes two origins and one Website protection licence. Origin capacity is pooled across the account. Each origin licence adds room for one more origin. The page shows how many origins you're using out of that total. When none are left, Add is turned off: add a licensed sentry or buy an origin licence on Billing first. An origin beyond the included ones uses an origin licence. Removing that origin returns its licence to your pool. See Licences.
Domains
Domains lists your hostnames. www is included automatically, so add the main domain or subdomain (example.com or app.example.com), not www.example.com. You can filter the list by Cluster, Origin, Type, HTTPS status (All statuses, or a single status) and Website protection rules (All, Log mode or Blocking).
Click Add domain to open its own page. Enter the Domain, and choose the Cluster, Origin server and Type (Generic, WordPress, Drupal or Joomla). WordPress turns on the WordPress rules when Website protection is on. If you type a leading www., the page notes the name that will be saved and that www is included automatically. Each domain has one origin server.
The Website protection button turns Website protection on or off for a domain (see Network protection vs Website protection). New domains with Website protection start in log mode. Each sentry includes one Website protection licence, pooled across the account; extra Website protection licences are added on Billing. The page shows how many domains with Website protection you're using out of that total. When there is no spare Website protection capacity, a greyed-out Off button explains on hover that you need to add a sentry or buy a Website protection licence, and Website protection cannot be turned on for another domain. If the account goes over its Website protection allowance, Website protection is turned off on the newest domains and the oldest keep Website protection. You can also select domains and use Turn on Website protection or Turn off Website protection. Turn on Website protection goes one domain at a time and stops at the first refusal, for example when there is no spare Website protection capacity. A domain in log mode shows a Log mode icon: The WAF and some Website protection rules on this domain record what they would block, but let those requests through. A domain with Under Attack on shows a red shield icon; the key under the table reads Under Attack is on. The table has no switch for Under Attack: turn it on or off from Settings › Captcha Protection, or from Under Attack for this domain on the domain's page. Account owners can select domains and use Switch to blocking or Switch to log mode, up to 500 at a time. For domains with Website protection off, the button in the PROXY column (PROXY protocol) sends the visitor's IP address to the origin on HTTPS. Set up the origin first: see PROXY protocol. Statuses show as icons, explained by a Key under the table (see HTTPS status). The origin and cluster appear under the domain name. The table also shows each domain's traffic for the last 60 minutes in Passthrough HTTPS / hr, Allowed/hr and Blocked/hr (see Domain traffic columns). Changes reach your sentries within about a minute.
HTTPS status
On Domains, statuses show as icons, with a Key under the table. The icons appear next to the Website protection button, except Not found on origin, which appears next to the domain name. The key uses these labels: Log mode, HTTPS inspected, HTTPS passed through, Certificate found, No valid certificate found, Origin agent not connected and Not found on origin. HTTPS inspected and HTTPS passed through only show when Website protection is on. The domain's own page still shows those two as badges, with the same wording.
The HTTPS statuses are decided in this order:
| What you see | What it means |
|---|---|
| HTTPS inspected | Website protection is on, a valid certificate has been accepted, and it has been delivered to a ready sentry with an approved HTTPS key. This stays inspected even if the origin agent later disconnects. |
| Origin agent not connected | There is no origin agent, or it is pending, removed or has not checked in. A domain that is already inspected keeps that status instead. |
| Certificate found | A currently valid covering certificate is stored, or was in the origin agent's list. This is also shown for domains with Website protection that are waiting on a sentry HTTPS key. It is not a promise that HTTPS inspection will succeed: a self-signed or untrusted certificate can still fail. |
| No valid certificate found | The certificate is expired, does not cover the domain, or was not found. |
| No HTTPS icon | The certificate has not been checked yet. |
In the HTTPS status filter, inspected domains are listed as Inspected. See HTTPS inspection.
Not found on origin
Not found on origin is only a flag. Protection continues. On Domains, it shows as an icon next to the domain name, and in the Key under the table. It appears when the origin agent is connected, its latest domain list was confirmed within 30 minutes, and neither the domain nor its www form is in that list. The tooltip reads: This domain wasn't in the latest domain list the origin agent sent from this origin server. It's still protected.
Domain traffic columns
The three traffic columns on Domains are:
- Passthrough HTTPS / hr — HTTPS that is passed through unopened to the origin: domains with Website protection off, or domains with Website protection without a usable certificate or sentry HTTPS key. The sentry counts the connection, but it does not see the pages, images or other files requested inside it. Under the figure, from N IP addresses is how many different visitor IP addresses opened those connections. Many addresses with only a few connections each usually means a scrape or a botnet, which per-IP rate limits don't stop.
- Allowed/hr — HTTP requests, and inspected HTTPS requests, that were let through without a block reason. This includes the site's own error responses, such as a 404 for a missing page.
- Blocked/hr — HTTP requests, and inspected HTTPS requests, that stillorigin blocked or answered with an error itself. The site's own error responses aren't counted as blocked.
Each column is a total for the last 60 minutes, added up across every sentry in the domain's cluster. It isn't a calendar hour, and it isn't an average. Traffic to www is counted on the main domain. The domain's own page and the dashboard show the last 24 hours instead.
Requests count individual requests, not people, so they aren't unique visitors.
Why connections and requests don't add up
Connections and requests count different things, so Passthrough HTTPS / hr isn't allowed plus blocked.
When HTTPS is passed through unopened, visitors to an HTTPS site show up only under Passthrough HTTPS / hr. The requests columns then show mostly plain HTTP on port 80, such as redirects to HTTPS and automated probes. That's why a row can show 264 connections and 29 requests.
When HTTPS is inspected, those requests count under Allowed/hr and Blocked/hr, exactly like HTTP. Passthrough HTTPS / hr for that domain drops to about zero, and its request figures rise. See HTTPS inspection.
A connection whose TLS handshake fails is not counted in any of these columns.
With Website protection off and Website protection on
- Website protection off: HTTPS on port 443 is passed through unopened and counted under Passthrough HTTPS / hr. Plain HTTP on port 80 is counted under the requests columns.
- Website protection on: HTTP on port 80 is inspected, and the requests columns include the Website protection rules: blocked paths, the WordPress rules, and HTTP 429 answers when a visitor goes over the Website protection request limit. In log mode, requests that the WAF (when it's on) and the path traversal and private file rules would have blocked count as allowed, not blocked. The WordPress hardening rules, the admin login lock if it's on, and rate limits still block, so those requests and HTTP 429 answers still count as blocked. For HTTPS, see HTTPS inspection. While HTTPS is passed through unopened (no usable certificate or sentry HTTPS key yet), connections still count under Passthrough HTTPS / hr. Once HTTPS is inspected, those requests count in the requests columns, and Passthrough HTTPS / hr drops to about zero.
See Network protection vs Website protection.
What the columns leave out
Packet-filter blocks aren't in any of the three columns. That includes rate-limit drops, temporary bans, deny-list drops and other firewall blocks. You'll find them on the dashboard's Firewall blocks chart, under Logs, and in the Identified traffic breakdown. A connection whose TLS handshake fails is not counted here either.
Domain traffic graphs
Click a domain's name to open its page. The header shows the domain name, its cluster and origin, Website protection on or Website protection off, and a Log mode badge when Website protection rules are in log mode. Edit domain is where you change settings. On domains with Website protection, an account owner can also Log in as domain owner, and Under Attack for this domain appears in the header.
After you change a domain, the page shows Applying your changes until every online sentry in its cluster has the change, which usually takes about a minute. It says how many sentries have it so far. If a sentry couldn't apply its latest settings, the notice says so and stays until that's fixed; your dashboard shows which sentry.
The page has five tabs: Overview, Traffic, Blocks, Protection and Countries.
Overview shows the last 24 hours: Allowed web requests, 24h, Blocked web requests, 24h, Passthrough HTTPS connections, 24h when there were passed-through connections, Browser checks passed, 24h and Browser checks failed, 24h when Captcha Protection applies, Top block reason, 24h, and Sentries reporting.
Traffic shows:
- Web requests — allowed and blocked HTTP requests per hour, with Website protection on or off, including inspected HTTPS requests;
- Passthrough HTTPS connections — HTTPS passed through unopened to the origin, per hour, next to Different IP addresses, the number of different visitor IP addresses behind them each hour. When HTTPS is inspected, this figure drops to about zero, and the requests chart rises;
- a map of requests by the visitor's country. Requests to this domain, by the visitor's country. Without Website protection, HTTPS connections go straight to the origin server and aren't counted. No IP addresses are stored or shown here.
- By sentry — how the traffic was shared between your sentries (see below).
Blocks shows Top block reasons and Top blocked IPs for the domain, with a link to Logs for an IP's full history, plus Identified traffic on this domain — traffic seen on this domain's hostname, grouped by type. Firewall blocks that have no hostname appear only in the cluster totals. Blocked-IP counts come from sampled events, so they show the pattern rather than exact totals.
When the domain is in log mode, the Blocks tab also shows Would have blocked, last 24 hours: a count, a breakdown by rule, and Recent requests recorded but let through. When the domain is set to blocking, the same section is headed Recorded but let through, last 24 hours and only appears when something was recorded: country rules in log mode, and rules recorded before you switched to blocking (see After you switch to blocking). Those requests count as allowed, not blocked, and they do not appear under Logs. That view is read-only. To ignore a WAF rule or switch between log mode and blocking, use Edit domain, under Website protection rules.
Protection is read-only: What's protecting this domain right now. To change a setting, use Edit. It lists Website protection, Website protection rules, WAF, WordPress rules (on WordPress domains, including the admin login lock), Captcha Protection, Rate limits and identified traffic, and HTTPS. Each setting has an Edit link to the matching section of Edit domain. If Website protection is off, Captcha Protection reads Captcha Protection needs Website protection. Turn on Website protection for this domain to use it. If the WAF is off in the cluster's settings, WAF shows: The WAF is off in this domain's cluster settings, so it isn't checking requests to this domain. See WAF rules and Captcha Protection.
Countries shows every country rule that applies to the domain: your network-wide countries, this domain's own, and any the domain owner added. Add countries for this domain only under This domain's countries. These need Website protection. See Country blocking.
To change a setting, click Edit domain. Settings on this page jumps to Cluster and origin server, Website protection rules, WAF, WordPress rules (WordPress domains only), Captcha Protection, Rate limits and identified traffic, HTTPS, and Domain owner access. Rate limits, identified traffic blocks and fake crawler blocking are set for the whole cluster, so they apply to every domain on it. From that section you can open Edit cluster, Identified traffic blocks for this cluster, or Captcha Protection for this cluster.
By sentry
By sentry lists each sentry that handled the domain in the last 24 hours, with its Requests allowed, Requests blocked and Passthrough HTTPS connections, with the number of different IP addresses under the connections. Share draws each sentry's allowed and blocked requests as a bar, scaled to the busiest sentry's requests. Passthrough HTTPS connections aren't included in the bar.
Deny list
Deny list blocks addresses at every sentry in a cluster. Choose the cluster and enter one or more IPv4 or IPv6 addresses (203.0.113.5, 2001:db8::5) or ranges (203.0.113.0/24, 2001:db8::/48), or AS numbers (AS132203, see Blocking a network by AS number), under Addresses, one per line, then an optional reason and whether they are Permanent. Every line gets the same reason and Permanent setting. Blank lines and anything after # are ignored, and you can add up to 2,000 at a time. Non-permanent entries expire after your deny expiry hours. Changes reach the sentries within about a minute.
After Add, the page says how many were added. Lines that weren't added are listed with the reason, such as not being an address or range, appearing earlier in your list, already being on the deny list, or being one of your origin servers or sentries. Those lines are left in the box so you can fix them and add them again.
Blocking a network by AS number
You can also block every network of an autonomous system (AS), such as a hosting company that scans your sites. Enter its AS number on a line of its own, for example AS132203, in the same Addresses box. It's added as one entry that shows the AS number, the network's name and how many ranges it covers, such as 1,234 ranges. Its ranges go to the sentries with the rest of the deny list. Like any deny entry, the sentry drops everything from them.
- An AS can't be added if any of its networks include one of your origin servers or sentries. Blocking it would take your sites offline.
- Verified crawler ranges, such as those of search engines, are never blocked by AS number. They're left out of the AS's ranges.
- A cluster can have up to 5,000 deny entries in all, counting each range of an AS. An AS that would take the cluster over that is refused, with how many entries are left.
- The ranges come from the IP to ASN Lite database by DB-IP (CC BY 4.0), updated each month. When it updates, each AS's ranges are updated too and reach the sentries within about a minute. An AS that has grown past the 5,000 limit keeps its previous ranges.
To stop blocking an AS, click Remove on its entry.
Some entries are added for you: addresses caught for excessive traffic, and addresses promoted from a temporary ban. For IPv6 these cover the address's whole /64. They show their reason in the list and can be removed like any other entry.
Click an IP in the list to Show location. It shows Country, City and Looked up, from the location stored for that address. If none is stored yet, it says No location is stored for this address yet. The location is approximate.
Deny expiry
Deny expiry, on Settings, is the number of hours a deny entry lasts when it isn't marked permanent. It applies to entries you add by hand and to automatic entries, such as addresses caught for excessive traffic or failed logins.
When it's set to 0, entries you add stay until you remove them, and automatic entries last 24 hours.
Logs
Logs has three tabs: Blocked IPs, Temporary bans and Auth fails. Individual events are kept for 3 days. The 7 and 30 day views show a daily summary per IP instead, kept for 90 days (see Time ranges).
Blocked IPs
Blocked IPs lists sampled blocks from your sentries: firewall blocks and refused HTTP requests on port 80. Refused requests include requests for a hostname you haven't added (unknown-host) and error responses from your origin on port 80. HTTPS that's passed straight through to your origin never appears here.
These are samples, not every blocked connection or request. Each address is recorded at most once per reason every 30 seconds, and firewall logging is rate limited. They won't match the blocked web requests total on the dashboard. Filter by cluster, sentry, reason and time.
Domain shows the domain a refused request was for. For a hostname you haven't added, such as an unknown-host block, it shows the hostname the request asked for, or the IP address if the visitor used that. Firewall blocks happen before any hostname is known, so they show -. Blocks recorded before this column was added show - too. Search also matches the domain, and the CSV export has a Domain column.
Temporary bans
Temporary bans lists the addresses each sentry is temporarily banning right now, with the time left. This is the live list, not the sampled log. Unban removes an address on every sentry in that cluster, usually within a minute. Promote adds it to the cluster's deny list as a timed or permanent entry.
Auth fails
Auth fails is for failed logins reported from your origin servers. Reporting failed logins from origin servers isn't available yet, so this tab is normally empty. See Auth fails.
Time ranges
On Blocked IPs and Auth fails:
- Last 24 hours and Last 3 days list individual events. Individual events are kept for 3 days.
- Last 7 days and Last 30 days show a daily summary: one row per IP, reason and day (UTC) with the number of events and when the IP was first and last seen that day. Summaries are kept for 90 days and updated hourly, so the latest hour may not be included yet. The summary is per cluster, so the sentry filter does not apply.
Search, sorting and pages
Search matches the start of an IP address (or the full address) and also the reason, sentry, path and country. Click a column heading to sort. Tables show 50 rows per page by default; choose 25, 50, 100 or 250 under the table and move between pages with Newer / Older (or Previous / Next).
Country and city are approximate location labels. They appear only after the address has been looked up; opening the page does not look addresses up.
Export CSV
Export CSV under each table downloads every row that matches the current tab, filters, search and sort, not just the page you are looking at, up to 100,000 rows. If there are more, the file ends with a note; narrow the time range or filters and export again. Times in the file are UTC.
What the reasons mean
Every reason is explained in Log reasons explained. Click a reason in any Logs table to jump to its explanation.