Skip to content

HiL Agent Status

When you run the ReSim HiL agent for your own hardware-in-the-loop workflows, you'll want to keep an eye on the whole fleet: which agents are online, what each is running, what's queued for every pool label, and how busy your agents have been. HiL Agent Status answers these operational questions across your organization.

It's available from two surfaces that read the same organization-wide data:

  • the resim CLI, under resim agents and resim pool-labels, and
  • the ReSim web app, under Analyze > HiL Agent Status.

Each section below gives the CLI command and where to find the same view in the app.

Viewing your agents

The fleet view lists every agent that has checked in, with its reported version, pool labels, last check-in time, and current status. An agent is active while it's checking in regularly and becomes inactive once it stops. Each agent also surfaces its most recent run.

Bash
resim agents list          # add --json for machine-readable output

In the app, the HiL Agents tab shows the same list, split into Active and Inactive sections.

Inspecting an agent's run history

To drill into a single agent and the tests it has run:

Bash
resim agents get --agent-id robot-arm-1

In the app, each agent has a detail view that lists every test that has been run on that agent.

Viewing the pool label queue

Pool labels group work by the kind of hardware it needs. To see what's running and waiting for each pool:

Bash
resim pool-labels queue                       # org-wide
resim pool-labels queue --completed-since-days 14

--completed-since-days sets how far back to include already-finished batches (1–30 days, default 7). In the app, the Pool label queue groups active and queued batches by pool label, and shows which agents serve each pool.

Measuring agent utilization

resim agents utilization returns a time-bucketed series showing how hard your agents are working over a window:

Bash
resim agents utilization                                              # whole org, last 7 days
resim agents utilization --agent-id robot-arm-1 --interval hour \
    --start-time 2026-06-04T00:00:00Z --end-time 2026-06-05T00:00:00Z

Each bucket reports the fraction of wall-clock time the agent spent running experiences (utilization), the fraction it was offline (heartbeat silent for over five minutes), and how many tests started. Omit --agent-id for the whole org or pass it for a single agent. --interval is hour or day (default day), and the window defaults to the last seven days. The summary also reports total tests run, the average and median queue wait (submission to execution start), and the top experiences by running time (--top-experiences, default 10). Utilization details are not yet available in the app.

Note

Utilization is measured from when a run starts to when it reports a terminal status. So a reading pinned at 100% could mean a single stuck run rather than a genuinely busy fleet.

Archiving an agent

When you decommission an agent it stays in the list as inactive. Archiving hides it from view:

Bash
resim agents archive --agent-id robot-arm-1    # aliases: remove, hide; add --yes to skip the prompt

In the app, click Remove on an inactive agent. Archiving only hides the agent — if that host ever checks in again, it reappears automatically. Active agents can't be archived.

Keeping agents up to date

ReSim tracks the latest released agent version. If an agent reports an older version, both the CLI and the web app flag it as out of date. Follow the installation steps to download and install the latest binary.