Create a Linear issue or Trello card for every incident.

If an outage only shows up as an alert email, it is easy to forget. Barkme creates a ticket for every incident, so someone owns it and there is a record.

Why alerts are not enough

An alert tells you something is broken right now. Once the site is back, the alert has done its job, and it sinks under the next hundred messages in the inbox or the Slack channel.

The outage itself has not been dealt with, though. Something caused it: a bad deploy, a full disk, a dependency that timed out. If nobody writes that down and looks into it, it happens again. A ticket in the tool your team already plans its work in gives the outage an owner, a place in the backlog and a history you can search later.

Linear and Trello

Connect Linear or Trello once. After that, every incident creates an issue or card with the monitor, the URL and the error. When the monitor recovers, Barkme adds a comment with the total downtime.

Barkme does not close the ticket. The site being back does not mean you know why it broke, so the ticket stays open until someone has checked.

Setting it up

The integrations sit at the end of the normal Barkme setup:

1. Verify your domain with one TXT record on the apex. It usually takes about two minutes and covers every subdomain.
2. Add your monitors. Each one checks a URL with an HTTP GET on an interval from 30 seconds to 5 minutes, and opens an incident after two failed checks in a row.
3. Connect Linear or Trello. You authorise Barkme in Linear or Trello, then pick where tickets go: a Linear team, or a Trello list. Only an owner or admin can set this up, because it is shared by the whole team.

Like every channel, Linear and Trello get a test when you connect them, so a wrong team or list shows up straight away, long before a real incident needs it.

What the ticket looks like

Say your checkout API starts failing at 02:14 and comes back at 02:31. Two failed checks in a row open an incident, backdated to the first failure, and a Linear issue appears in the team you chose. Its title is DOWN followed by the monitor name, and the description lists:

• Monitor and URL, so you know what failed.
• Endpoint, the host that was checked.
• Detected, the time of the first failed check in your team's timezone.
• Error and Failed at, when the request failed before a response came back, such as a timeout at the connection stage or a TLS handshake error.
• Probable cause, when the error matches a known pattern, such as a firewall blocking the prober.

At 02:31 the next check passes. The incident closes, and Barkme comments on the same issue: the monitor is responding again, the time it recovered, and 17 minutes of downtime. The comment also says the issue was left open on purpose, to be closed once the cause has been looked into.

A Trello card works the same way. The card is created in the list you picked with the same details, and the recovery goes on the card as a comment.

One ticket per incident

A monitor has only one open incident at a time. However long an outage lasts, it creates one issue or card, and the recovery is a comment on that same ticket. You will not find twenty cards for one bad night.

Linear and Trello only handle incidents. SSL and domain expiry warnings and blocklist alerts go to email, Slack, Telegram and the webhook. If you want those as well, see SSL and domain expiry.

Webhook for other tools

For your own tools, a webhook sends each event as JSON, signed with HMAC-SHA256 in the X-Barkme-Signature header so you can check it came from Barkme.

The signature covers the timestamp and the raw body, using a signing secret shown on the channel. Recompute it on your side and drop any request that does not match. The webhook receives incident events (incident.opened and incident.resolved, with the monitor name, URL, start and resolve times, error and probable cause) as well as expiry and blocklist alerts. The exact payload is shown on the channel page when you connect it.

There is no built-in Jira, GitHub Issues or PagerDuty integration. The webhook is how you reach them: point it at a small endpoint of your own that creates the ticket wherever your team works.

Tested when you connect it

Every channel sends a test alert when you connect it, so you find out right away if something is set up wrong.

You can also mute a channel any time without deleting it, for example while you move boards around, and turn it back on when you are done.

Plan

Linear, Trello and the webhook are on Pro. Slack and Telegram are on Founder, and email is on every plan.

Pro is $39 a month and also removes the monitor cap, covers up to 50 verified domains and has no seat limit, which suits an engineering team that wants every outage tracked. If you only need to know when something is down, Founder at $19 a month with Slack alerts may be enough. The pricing page compares all three plans, and our post on choosing uptime alert channels goes through which channel suits which job.