Alerts and thresholds

Raccoon emails every member of the organization when something needs attention. Each member chooses which emails they get, per environment. For chat tools and paging, use webhooks.

Emails#

Email	Sent when
New issues	An exception issue opens.
Reopened issues	A resolved issue happens again. Reopening an issue by hand doesn't send it.
Assigned to me	Someone assigns an issue to you. Only the assignee gets it.
Thresholds exceeded	A route, job, command or scheduled task goes over its threshold.
Stopped sending events	An environment with a silence alert goes quiet, and again when events resume.
Weekly report	Every Monday. See Weekly report.
Weekly digest (Pro)	Every Monday. See Assistant and weekly digest.

Resolving or ignoring an issue doesn't send an email. Ignored issues send nothing more.

Each email's subject starts with the application and environment, such as [Shop (Production)] New issue #12: TypeError.

Choosing your emails#

Open Notifications from the menu under your name. Every email is on until you turn it off. Your choices only affect you, not the rest of your team. They're set per environment, so you can keep Production alerts and turn off Staging.

Weekly report#

Every Monday at 07:00 UTC, each member gets one email per organization covering the previous Monday to Sunday, on every plan:

• Exceptions: the total, the users they affected and the change from the week before, then per environment the exceptions, users, and issues opened and resolved.
• Requests: the total and the change from the week before, then per environment the requests by status: 1/2/3XX, 4XX and 5XX.
• Event allowance: how much of this month's allowance is used.

It covers the environments that sent events in the last two weeks. Turn off Weekly report for an environment under Notifications to leave it out of your report. The link at the bottom of the email turns it off for the whole organization.

Thresholds#

A threshold is the duration you consider too slow for a route, job, command or scheduled task. Set them under Settings → Thresholds. They apply to every environment of the application.

1. Pick the route, job, command or scheduled task. The list shows what ran in the last 30 days.
2. Enter a duration in milliseconds.

Every minute, Raccoon computes each item's 95th percentile duration over the previous 5 minutes. When that's above the threshold, it opens a performance issue for that item and emails the "Thresholds exceeded" alert, which states the p95 and the threshold.

While the issue stays open, Raccoon doesn't email again, and only counts occurrences. Resolve the issue once the fix is out. If the item gets slow again after that, the issue reopens and you get another email.

Thresholds also show in the dashboard:

• Duration charts draw the threshold as a dashed line.
• Jobs, commands and scheduled tasks show durations over their threshold in red, and their pages have a ≥ Threshold filter.
• The Dashboard's Routes card counts routes over their thresholds.

Members can see thresholds. Admins and owners can add, change and delete them.

Silence alerts#

A silence alert emails you when an environment stops sending events, such as when an agent crashes, a token is rotated or a deploy breaks the setup. Turn it on per environment under Settings → Environments, with Alert when no events for set to 15 minutes, 1 hour, 6 hours or 1 day.

Raccoon checks every five minutes. It sends one email when the environment goes quiet, and the environment's row shows "Silence alert sent". When events arrive again, members get an "Events are arriving again" email. Environments that have never sent an event aren't checked.

Silence alerts also go to webhooks subscribed to environment.silent and environment.recovered.

Pick a time longer than your quietest normal stretch. A staging app that nobody uses at night may need 6 hours or a day.