executor_idstring uuid required
ssh_connect_cmdstring required
cpu_countinteger required
ram_totalinteger required
is_favoriteboolean nullable required
updated_atstring date-time nullable required
created_atstring date-time nullable required
removal_scheduled_atstring date-time nullable required
termination_hoursinteger nullable required
jupyter_installation_status'INSTALLING' | 'SUCCESS' | 'FAILED' required
jupyter_urlstring nullable required
enable_jupyterboolean nullable required
enable_volume_encryptionboolean nullable required
volume_encryption_status'ENABLED' | 'UNSUPPORTED_IMAGE' | 'DISABLED' | 'FAILED' required
pending_pricenumber nullable
Hourly USD price this pod moves to at price_change_effective_at: the node's listed price is now above what this pod pays. null when no change is pending.
price_change_effective_atstring date-time nullable
UTC time the pending price takes effect. Null when no change is pending or while its notification e-mail is held. Notices are limited to one per pod per 24 hours.
price_change_notice_hoursinteger
Hours a pending raise waits after the renter is e-mailed about it before it reaches the pod (24 on a Spot node).
cluster_idstring uuid nullable
cluster_node_indexinteger nullable
created_by_api_key_idstring uuid nullable
api_key_namestring nullable
workspace_idstring uuid nullable
volume_limit_gbinteger nullable
Size in GB of the pod's local /root volume. null on a pod created without a storage limit.
spend_to_datenumber
USD billed to this rental so far, summed from the billing ledger — not price × elapsed time. The ledger accrues every 5 minutes and settles on deletion, so this lags real time by at most one accrual interval; 0.0 until the first charge is booked.
billed_untilstring date-time nullable
UTC time up to which spend_to_date is settled: the ledger's last accrual tick for this rental. null until the first charge is booked (a pod in its first minutes, or one that never became billable).
skip_add_agent_ssh_keyboolean nullable
estimated_ready_secondsinteger nullable
Seconds left of the time-to-ready estimate made when the rent was accepted; 0 once it has run over. Only while the pod is PENDING or REBOOT_PENDING.
eta_basis'executor_history' | 'class_median' | 'cold_pull_estimate' | 'fleet_default' nullable
What the estimate was made from: this node's recent creates (executor_history), its GPU class (class_median) or the fleet default (fleet_default); cold_pull_estimate when an image pull the node does not have cached was added on top.
phase'queued' | 'preparing node' | 'connecting to node' | 'pulling image' | 'creating volume' | 'starting container' | 'mounting encrypted volume' | 'configuring ssh' nullable
The step the create has reached, read from the validator's log lines.
rental_tierstring nullable
starting_ssh_connect_cmdstring nullable
The SSH command the pod will answer on, set while it is PENDING or REBOOT_PENDING once its SSH port is known, before the API reports the pod RUNNING. Null while the port is not known yet, and in every other status (use ssh_connect_cmd there). Older servers do not send this field.
cluster_overlay_ipstring nullable required
This node's address on the cluster's private overlay — where its peers reach it.