Find out a client site is down before the client does.

If you look after websites for clients, you want to hear about problems before they do. Barkme watches every client site and tells you when something breaks.

The problem with finding out from the client

Most agencies have no monitoring on client sites at all, or a free checker set up years ago that emails an address nobody reads. The first sign of trouble is a phone call. By then the site has been down for a while, the client has heard about it from a customer, and the five-minute fix comes with an awkward conversation about why you did not know.

Outages are not the only thing that goes wrong. Certificates expire. Domains lapse because the renewal email went to someone who left. A client's domain ends up on a mail blocklist and their invoices stop arriving. Each of these is easy to fix and hard to notice, and each one lands on you.

Add each client's domain

Verify a client's domain once by adding a DNS TXT record. That covers every subdomain, so you can monitor the client's site, shop and API without asking them for anything else. Each monitor requests a URL from outside, the way a visitor would. A 2xx or 3xx response counts as up.

Setting up a client

The setup is the same for every client, and it takes a few minutes each:

1. Verify the domain. Add a TXT record on the apex with host @ and the value Barkme shows you. It usually takes about two minutes. If the client manages their own DNS, send them the record.
2. Create a project for the client. Projects keep each client's monitors together, and each project can have its own status page.
3. Add monitors. One for the homepage, one for the shop checkout, one for the API or the booking system, whatever matters to that client. Pick the interval and timeout per monitor.
4. Connect your channels once. Channels are not tied to a client, so once Slack is connected, every client's alerts arrive there. Each channel sends a test alert when you connect it.
5. Create the status page and send the client the link.

Alerts only for real outages

Barkme checks again before it alerts you. By default an incident opens only after two failed checks in a row, so a single slow response does not wake anyone up. The incident starts from the first failure, so the downtime you report to the client is accurate. It closes on the first successful check.

On Founder and Pro, sites are checked as often as every 30 seconds, and alerts can go to Slack or Telegram as well as email.

A lot of client sites run behind a firewall plugin or a host's bot protection. If one of them starts blocking the prober, the alert names that as the probable cause, and you can allow the Barkme prober IP shown in your dashboard. For more on this, see why uptime monitors send false alarms.

What the alert looks like in Slack

Say a client's homepage starts timing out at 16:00. The first check fails, Barkme retries after two seconds, and the next check fails too. At 16:01 a message lands in your team's Slack channel, headed DOWN with the monitor's name. It lists the URL, the host, when the first failure was detected, the error and the stage it failed at, such as the connection or the TLS handshake.

Whoever is free picks it up. The incident stays open while the site is down, and Barkme does not send another alert for the same outage. When the next check passes at 16:06, a second message says the site has recovered, with the time it came back. The incident record keeps the five minutes of downtime, backdated to 16:00, so the number you give the client is the real one.

If you want every client outage to become a ticket as well, Pro adds Linear and Trello. See incidents as Linear issues and Trello cards.

Expiry and blocklist warnings

Barkme also checks every verified domain's SSL certificate and registration once a day, and warns you 30, 14, 7 and 1 days before either expires. A daily check tells you when a client domain is put on a security filter or a mail blocklist, and again when it is removed.

The blocklist check runs on two tracks. The web track compares the domain's DNS answers from security-filtering resolvers (Cloudflare and AdGuard, plus Quad9's threat feeds) with clean resolvers, which catches the day a security filter starts blocking the site. The mail track looks the domain up on Spamhaus DBL, SURBL and URIBL, the lists that decide whether the client's email lands.

For a quick look at a prospect's site before they sign, the free domain health check shows certificate and registration expiry, DNS and blocklist status with no account needed.

A status page for each client

Each project can have a public status page that shows what is monitored and any recent incidents. On paid plans it can use the client's own domain, so the client can check it without calling you.

Monitor URLs are never shown, you choose a display name for each monitor, and the page shows 90 days of uptime. You can add the client's logo and accent colour. More on this in client status pages.

Your team

Founder includes five seats, and Pro has no seat limit. An owner manages billing, an admin can invite people and set up the shared channels, and a member works in the dashboard without managing the team. Slack, Linear, Trello and the webhook are shared by the whole team. Email and Telegram are personal, so each person connects their own and their alerts go only to them.

Which plan fits

Founder ($19/mo) fits a freelancer or a small studio with up to five client domains: 30 monitors, 30-second checks, Slack and Telegram, five seats and a status page on each client's own domain.

If you look after more than five client domains, Pro ($39/mo) is the plan. It covers up to 50 verified domains with unlimited monitors and seats, and adds Linear, Trello and a signed webhook. The pricing page has the full comparison, and if you are moving from another tool, see how Barkme compares with UptimeRobot.