Skip to main content

Nodes and capacity

Lesson 10 of 11 · 2 min 29 s. Episode 10 of the Jaah tutorial series: give your team enough capacity, and see that it is healthy. Start a cloud machine from your own AWS account and watch it come online, see what a machine needs before it runs work, add places to work so tasks run side by side instead of taking turns, and read a machine's health and metrics.

Before you start​

  • An AWS credential set up for cloud machines. Where its machines run is set once, under Settings.
  • A project whose work already runs on one of your machines. In the video, TimeSaver (a time-tracking app for small teams) adds capacity to the TimeSaver app project.

Steps​

  1. Open Directory and go to the Nodes tab. Each line is one machine. The green dot means it's online: connected and ready for work.
  2. Press Add node and choose Cloud. It starts a machine for you in your own AWS account, and it's billed there; you can also register one of your own. Under Name, give it a name. The AWS credential it starts from is chosen already when it's the only one set up for cloud machines. Press Create.
  3. Watch the new machine appear on the list while it starts, which takes a few minutes. When its dot turns green, it's online, and it shows what it has to offer, like any other machine.
  4. Go to the Model Accounts tab. A machine runs work once the model account your agents use is signed in on it, a one-time step for a person. Each account shows which agent it's for and how much of its quota is left.
  5. Open the project from Projects and go to its Worktrees tab. Each place to work lets one of the project's tasks run at a time, so a project allowed only one has its tasks take turns. Raise Pool max to the number of places you want, and press Save.
  6. On Directory's Worktrees tab, press Add slots. Under Node, pick the machine that already runs the project's work. Under Project, pick the project. Under Worktrees to add, type how many more. Press Add worktrees.
  7. Go back to Tasks and pick the project. Its requests no longer wait in a single line. Two are worked on at the same time, each in its own place to work, and the third starts as soon as one finishes. A small machine runs two at a time; a bigger one runs more.
  8. To see how a machine is coping, open it from Nodes and go to Health & Metrics. At the top is the verdict. Below it are its CPU and memory over time, so you can tell when it's time for a bigger machine or another one.

What you learned​

  • Machines are listed under Directory, on Nodes. Add node with Cloud starts one in your own AWS account.
  • A machine runs work only once the model account your agents use is signed in on it.
  • Capacity is places to work: raise the project's Pool max, then add slots on the machine that runs its work.
  • Health & Metrics shows whether a machine is healthy and how busy it is.

Next​

Next lesson: Access control and audit. Previous: Adding a project.

Related: How nodes work · Add a node · How worktrees work