Questions.

Websites and APIs over HTTP, TCP ports for VPNs and gateways, ICMP ping, DNS records, TLS certificates, domain registrations, heartbeats for scheduled jobs, anything in your network through a Beacon on a box there, mail health, the names under a domain looked at daily for takeover risk, an office or ISP link pinged externally and confirmed internally, a vendor's own status page, and certificates or statuses reported from places hostnag can't reach on its own. Fourteen monitor types in all.

Pick from six presets: 1, 5, 15, 60, 360 or 1440 minutes. Free floors at 5 minutes for uptime-style monitors and 1440 for expiry ones; Pro drops that to 1 and 60, Business to 1 and 15.

Five locations: US East, US West, Frankfurt, Tokyo and Sydney. Free watches from one of them, Pro from up to three, Business from all five.

Yes. Ping monitors send ICMP echo requests from five places of our own (US East, US West, Frankfurt, Tokyo, Sydney), several packets per run, and record replies, loss and round-trip time per place. Nothing is installed on your side.

Yes, with a Beacon, a plain script that one line installs on a box in the network. A website, port, DNS, ping or certificate check that picks the beacon under Checked by is run from there and judged by hostnag with the same rules; a private address is accepted for a check beacons alone run. Without a beacon, private and loopback addresses are refused as both targets and resolvers, on purpose; a host can still report its own status through a status report URL.

15 seconds for uptime-style monitors like HTTP, TCP, DNS and links. Expiry monitors, certificates and domain registrations, get 30 seconds, since RDAP and TLS handshakes can run slower.

Yes, from UptimeRobot, Better Stack, Pingdom, Uptime Kuma or a CSV file. hostnag previews every candidate first and creates monitors up to your plan's limit; an API key or token is used once to fetch the list and is never stored.

No. A failure needs to be confirmed, either a second location seeing it or a second consecutive run, before hostnag opens an alert.

30, 14, 7, 3 and 1 days out by default, and you can change them per group. The final window repeats daily, so a certificate or domain close to lapsing never falls out of view.

Nine kinds: email, Slack, Discord, Teams, Telegram and webhook on every plan, plus PagerDuty and Opsgenie on Pro, plus SMS on Business (300 texts a month). Set them up once for the org and route them per group.

Yes, both for a whole group and for a single monitor, and you can also set quiet hours on your own timezone. Alerts that go unacknowledged escalate through up to three tiers, each after its own number of minutes, until someone acknowledges from inside the alert.

Yes. A recovery notice goes out the moment a monitor clears and says how long it was down, so you do not have to open the app to find out.

JSON at version 4, with the group and the details of what happened, signed with x-hostnag-* headers so you can verify it came from hostnag. Any breaking change bumps the version number.

Uptime, response time and expiry for the monitors you choose, and a day by day health strip for mail, certificate and domain records, under groups you name or your own sub-zones, updated at most a minute behind live, never the actual targets or error strings. Incidents are written by you with an update ladder and a postmortem, or raised by the checks on their own, and listed day by day. The layout is yours: a headline band or a plain card, groups as cards, simple or detailed figures, response time a click away. Visitors subscribe by email, read an Atom feed or a Statuspage-shaped JSON, and every page and monitor gets its own SVG badge. Pro and Business can put a page behind a passphrase.

Yes. Tools, in the app, runs any check once: paste an address and the sensible checks are ticked for you, and any result becomes a monitor in one click. For a client review or an audit, gather checks into an assessment with notes and your own status per check, run them again for before and after, and print, share or export the report.

Yes, one per group, emailed to owners and admins on the 1st. You can set an SLA target for it; Pro keeps 3 months of reports, Business keeps 12.

20 monitors in three groups, three seats, checks every 5 minutes from one location, alerts by email, Slack, Discord, Teams, Telegram and webhook, one status page, and 30 days of history. It costs nothing and needs no card.

Every new organization starts with 14 days of Pro, no card: every minute from up to three places, 100 monitors, PagerDuty and Opsgenie, and the rest of Pro. When it ends the organization moves to Free on its own, and anything beyond Free's limits is paused, never deleted. Start a paid plan from the billing page at any point and the trial ends there.

Add monitor packs: 100 more monitors each for $15 a month ($150 a year), from the billing page. Pro takes up to 4 packs (500 monitors in all); Business starts at 500 and takes as many packs as you need. Beacon packs work the same way: 10 more beacons each for $10 a month ($100 a year), up to 9 on Pro (100 beacons in all), as many as you need on Business.

Yes: ten months for twelve. Pro is $190 a year, Business $790, a monitor pack $150, a beacon pack $100. Switch between monthly and yearly from the billing page whenever you like; the difference is prorated.

From the billing page, in every direction: up, down, monthly to yearly and back, packs on or off. Your first paid plan goes through Stripe Checkout. Once you are subscribed a bigger plan or more packs apply at once and charge the difference for the rest of the period; a smaller plan or fewer packs are credited on the next invoice.

Monitors over the new plan's limit are paused, not deleted. Turn them back on, or move up again, and nothing about them was lost. Moving to Free takes effect when the period ends.

Yes, from the billing page: choose Free and your plan runs until the period ends, then the organization moves to Free. There's no minimum term and no call required.

No. The trial and the Free plan need no card; you only enter one when you start a paid plan.

Three seats on Free, ten on Pro, as many as you need on Business. Every seat can be an owner, admin, member or viewer.

Yes, an authenticator app plus one-time recovery codes. An org can require it for every owner and admin, so the policy isn't left to individual habit.

Yes, a list of every active session. Sign one out on its own, or end everything else at once from a single device.

In Supabase Vault; hostnag's own database only ever holds a reference to them, never the value. Custom HTTP headers on a monitor are the one exception, stored in plaintext, and the form tells you so before you type one in.

The data lives on Supabase in us-east-1 and the app on Vercel. The checks run on small Fly.io machines in the locations you've picked and dial out from there.

30 days on Free, 90 on Pro, 365 on Business. Older results roll off.

A plain script, beacon.sh or beacon.ps1, on a box in your network: the internal counterpart of our check locations. It runs the checks that picked it (websites, ports, DNS, ping, TLS certificates) from the internal side, feeds the links, statuses and certificates that picked it, and reports in every few minutes. hostnag judges what it observed with the same rules as a live check, and the box shows as one more cell on the board. Beacons stand in groups; several beacons in a group are failover, and any one of them counts. When a box goes silent, you are told. You can read every line before you run it.

On the Beacons page, add a group (an office, a plant, a rack; it alerts as one of your zones) and a beacon in it, then run the one line on the beacon's page: it fetches the script, writes its config and schedules it every minute under your own scheduler (a systemd timer or cron, launchd, or a hidden Task Scheduler task). Any check picks the groups or beacons that run it under Checked by (a zone's only group is picked for you); the beacon picks it up on its next check-in. A status or certificate line needs its command or path in hostnag.conf on the box, and the beacon's page shows the line. Add allow lines there to fence what the beacon may reach.

Run beacon.sh uninstall (or beacon.ps1 uninstall). It removes the schedule, the script, the config and the state directory, and lists each. Then delete the beacon on its page, or hostnag reports it as silent.

Run it by hand with --verbose (-Verbose on Windows) and it prints one line per check it ran or fed (a reading or an error word, never a page body or a certificate) plus the check-in result; beacon.sh status prints the paths, the schedule, the last check-in and the last run, which is what a support reply asks for. The beacon's page shows the same, line by line.