A project groups the endpoints of one product or team, for example Checkout with its website, API, database and cache. Projects give you a reliability target with an error budget, a section on the status board, analytics over time, and a way to send each team only its own alerts.

1. Create the project#

Open Projects and click New project.
Give it a name, a colour and a priority.
Save. It starts with no endpoints and shows No endpoints.

2. Add its endpoints#

Click the project card, then add an endpoint from its panel, or
pick the project when you add or edit an endpoint on Services → Endpoints.

Endpoints that belong to no project are listed under Orphan Endpoints on the Projects page, where you can move them in.

A good first project has the public entry point (the website or API), the services it calls, and the database and cache behind them. With their metrics linked (see Connected monitoring), Pulsimo can also map the dependencies between them, so an outage of the database is one alarm about the database, not one per service.

3. Set the reliability target#

Every project has an uptime target, 99.5% unless you change it. The target sets the error budget: at 99.5%, half a percent of checks may fail. The project card shows how much of that budget is used; above 100%, the project is below its target.

To change it, open Project Analytics, click the project, and use Change target (99%, 99.5%, 99.9%, 99.95% or any value from 50 to 99.999). Members and administrators can change it.

Target	Failed checks allowed	About this much downtime in 30 days
99%	1 in 100	7 h 12 min
99.5%	1 in 200	3 h 36 min
99.9%	1 in 1,000	43 min
99.95%	1 in 2,000	22 min

Uptime counts checks, not minutes, and each endpoint counts the same in the project's average whatever its interval.

4. Send its alerts to the right people#

On Alerting → Channels, edit a channel and pick the project under Projects: that channel then only gets alarms about this project (alarms that belong to no project still reach every channel). A common setup is one channel per team plus one on-call channel set to Critical only. See Alarms and rules.

5. Plan maintenance#

Before a planned upgrade, use Alerting → Plan maintenance and choose the project (or some of its endpoints). A window can run once or every week, for example Sunday 02:00 for the weekly backup. Alarms during the window are recorded but held, and sent only if the problem is still there when the window ends. See Alarms and rules.

6. Watch how it does#

Page	Answers
Services → Overview	What is up or down right now, grouped by project
Projects	How each project does against its target over 7, 30 or 90 days
Project Analytics	Uptime per day, incidents by severity, time to acknowledge and resolve, weakest endpoints, with CSV export

All three are described in Projects and overview.