Glossary
The words Harthelm uses, each in one line.
Last updated 2026-10-19
| Term | Meaning |
|---|---|
| Always running | A workspace or service that never goes to sleep (availability always). The opposite of sleeps when idle. |
| API key | A token hart_<id>_<secret> that acts as you in hart and the REST API. |
| Base image | The container image a workspace's editor and terminals run in. The installation offers one or more. |
| Browser host | A RabbitMQ broker's or S3 store's own host ms-<n>.<workspace domain>, behind the team's sign-in or (S3 only) public. Since 1.4.0. See Managed services › From a browser. |
| Blocker | A reason a Sync cannot run yet, shown before you confirm it. See Sync blockers. |
| Console | Harthelm's web interface, for example https://console.example.com. |
| Context | A saved installation, API key and default team and workspace of hart. See hart ctx. |
| Custom domain | A domain of your own that serves a team's workspaces, verified with a TXT record. See Custom domains. |
| Dev URL | The address of a workspace's landscape, https://<id>-3000.<workspace domain>; its routes lead to the services. |
| Drain (night drain) | An installation can take nodes out of service at night; the workspaces on them go to sleep and wake elsewhere when opened. |
| Editor (IDE) | VS Code in the browser, one per workspace. |
| Environment variable | A name and value set for a workspace (all its services) or in one service's env. See Workspace environment. |
| Forward service | A service of type forward: no pods, the router passes its routes on to another address. |
| harthelm.yml | The file in /data/app that describes the landscape. See the reference. |
| hart | Harthelm's command line. See the reference. |
| Health check | How Harthelm tells that a service answers: an HTTP request or an open port. |
| Landscape | The services of one workspace, as harthelm.yml describes them and a Sync applied them. See Landscapes. |
| Managed service | A database or other service that a provider runs for you, such as PostgreSQL. See Managed services. |
| Needs | The services a service depends on: they wake and start first. |
| Sleeps when idle | A workspace or service that sleeps after its idle time and wakes when someone opens it (availability on-demand). |
| Operation | A long-running action of the API, such as creating a workspace, with a state you can follow. |
| Organization | The top level of who owns what: it holds teams and has owners. See Organizations and teams. |
| Plan | A size: CPU, memory and storage of a workspace, or CPU, memory and replicas of a service. See Limits › Plans. |
| Prepare | The stage that installs dependencies and builds, before the services start. |
| Problem type | The machine-readable kind of an API error. See Problem types. |
| Profile | A variant of the landscape in harthelm.<profile>.yml. See Profiles. |
| Public port | A service port with its own address, https://<id>-<port>.<workspace domain>. |
| Quota | The CPU, memory and storage a team may use in total. |
| Replica | One pod of a service; a service can run several. |
| Route | A path of the dev URL or of a custom domain that leads to a service's port. |
| Router | The part of Harthelm that serves the workspace addresses, signs people in and wakes sleeping workspaces. |
| Run | The stage that starts the services' steps; also the button that starts it. |
| Run switch | Whether a service should run (started) or not (stopped). Run and Start turn it on, Stop turns it off. |
| Secret | A value kept in the workspace's vault, readable only by the services that reference it. See Secrets. |
| Service | One part of a landscape: steps, container, managed or forward. See Services. |
| Sleep | A sleeping workspace or service has no pods; its files are kept. Opening it wakes it. See Sleep and wake. |
| Stage | Prepare, test or run: the steps Harthelm runs in a workspace, in that order. |
| Step | One shell command of a stage or of a steps service, run with bash -lc in /data/app. |
| Sync | Applies harthelm.yml to the cluster: creates, changes and removes services. You review the plan first. |
| Team | A group of members that share workspaces, a quota and custom domains. |
| Teardown | Removes every service of a landscape; code, variables and secrets stay, and a Sync builds it again. |
| Test | The stage that runs your tests, like prepare. |
| Wake | Starting a sleeping workspace or service again, by opening it or with hart ws wake. |
| Workspace | Your code on a volume of its own, an editor, terminals and the services of its landscape. See Workspaces. |
| Workspace domain | The domain under which an installation serves workspace addresses, for example ws.example.com. |