Changelog

An integration can now be restricted to some of your organisation's assets, for example to give a partner read access to one asset: pass asset_ids when you create it. A restricted integration is read-only and must hold exactly asset.read. The list shows asset_ids for each restricted integration and omits it for one that reaches every asset.

FCR now has backup markets too. Set an asset's FCR backup-market default with PUT /v1/assets/{asset_id}/backup_markets/fcr/default, and override it for a window with a bid under /v1/assets/{asset_id}/backup_markets/fcr/bids. They work as the aFRR endpoints do, but take no product, since FCR is symmetric.

Every FCR, aFRR capacity and aFRR energy result now names the bid it was built from, in bid_id. Use it to match a result to its bid: start_time, end_time and product alone are ambiguous when several bids share an interval and product.

A fault can now hold an asset to the power it still has instead of taking it out entirely: pass power_limit, the upper and lower bounds in MW. amend narrows an active hold after a further loss, and unavailability_hours sets how long the hold lasts, from 1 to 720 hours, 8 by default. The message draft takes the same fields, so the reviewed text describes what is declared.

Read one TSO pool's alarm history over a time range: the history counterpart of GET /v1/alarms/active/fleet, under alarm.read. Like every pool alarm, it is restricted to EBX.

For an open alarm, the triage context reports what the asset can still do: its available power and energy per direction against nameplate, the bounds a partial fault would declare, and the other assets in its pool with their alarms. It is read-only and needs alarm.write.

Onboardings now say whether they are a transition, in is_transition: an asset already live elsewhere, onboarded to produce its PQ application. A transition links the existing asset, starts with the hardware, software and monitoring-test gates skipped, and sends no invitations. Only EBX can create one, and for every other onboarding is_transition is false.

The RfI can record PQ list selections, such as the TSO, DSO or balance group, in pq_selections, picked from the PQ portal's catalogues. Read a catalogue with GET /v1/onboardings/{onboarding_id}/rfi-catalogues/{catalogue}, or under /v1/onboarding-tasks when you hold the task.

Promoting a reviewed draft to the official PQ application now starts with POST /v1/onboardings/{onboarding_id}/pq-official/confirmation, and the official sync carries its confirmation_id. When the sync review reports hierarchy_repair, send confirm_hierarchy_repair as true on both the confirmation and the sync, and leave it false otherwise; a mismatch is refused with 409, and only EBX can confirm a hierarchy repair. Test runs and sync reviews report input_issues, and a sync review also reports the DSO and any resources missing from the PQ portal.

An asset's FCR and aFRR pool membership has its own endpoint. GET /v1/assets/{asset_id}/tso-pool reads it under tso.read, PATCH switches either pool under tso.write, and both scopes can now be granted to integrations. Out of a pool, the asset's share of that market is covered by the other assets and none of its data for that market is sent to the TSO; bids are still accepted. On the asset itself, fcr_enabled and afrr_enabled are now read-only and deprecated, tso_connected is deprecated, and all three are ignored on create and update. They will be removed in v2.

Setting, shortening or deleting a power limit, energy limit or unavailability directly now needs restrictions.write_direct alone; asset.write is no longer required as well. A user or integration that holds restrictions.write_direct without asset.write was refused before and is accepted now.

Onboarding documents now download as an attachment under the file name they were uploaded with. Pass inline=true to preview a PDF, PNG, JPEG or WebP in the browser instead; any other type is refused with 409.

Set one PQ-workbook field of an onboarding's RfI, such as the planned CoD, without sending the whole RfI. Every other value stays as stored, so the call cannot overwrite an edit saved in between. It needs onboarding.write, and the value is validated and attributed as in a full RfI update.

Lists the same onboardings as GET /v1/onboardings, with the same filters and visibility, but only where each one stands: its asset, overall status, current gate, and each gate's status and deadline in process order. Use it to find an onboarding, then fetch the full record by id. It needs onboarding.read.

Whoever holds an onboarding's EMS task can now set the EMS connection with PATCH /v1/onboarding-tasks/{onboarding_id}/ems-network, which changes only the network fields it is sent. The task view carries asset_id and ems_network, and the RfI network configuration gains gateway and subnet_mask, empty until set.

Each RfI field's metadata now records who last changed it, in last_changed_by: a user_id, the email where known, and kind, user or system. It is null until a change has been attributed.

Your own profile now carries organization_name, your organisation's registered company name or null, and analytics_allowed, whether EBX's usage analytics may record you. It is false for integrations and for anyone who has opted out.

Prequalification with the TSO is now prepared in two workspaces, training and official. Choose the target and a contract per reserve type, run test submissions and curve checks in training, review each sync before it updates a draft, and promote a reviewed draft to the official application. Reading needs onboarding.read and every change onboarding.write. The PQ profile also gains delivery_duration_caps_hours, an optional cap per PQ product on the delivery duration the application requests.

Two reports for an asset, under report.read. The daily report covers one UTC day, today by default: aFRR activations with their fill rate, dispatch overrides, unavailabilities, alarms and the settled aFRR capacity revenue, which reads null, not zero, until a settlement result exists. The performance timeline gives activation metrics and settled capacity revenue per fifteen_minutes, hourly or daily bucket over [start_time, end_time): the last 24 hours by default, and at most 90 days.

EBX can now share notes and open points with the asset owner during an onboarding. Onboarding responses carry them in notes and shared_open_points, and the asset owner's task view in shared_notes and shared_open_points, each open point with its owner, due date, decision and comment thread. The asset owner can reply on an open point until EBX closes it with a decision; after that, a reply is refused with 409.

An onboarding can now be aborted, with a reason and optionally a category; the reason is recorded in its history. An abort is final. It is refused with 409 while a Betriebsfahrt session drives the asset or a PQ request is in flight, and for an onboarding that is already completed or aborted.

A block can record its cause: category, one of the open points' six topics, and waiting_on, the party that has to act. Both are optional, so a block without them is accepted as before.

Betriebsfahrt sessions now say when the plan is handed over for execution (handoff_at) and whether the drive asks for manual action, such as zeroing the battery (manual_action_required; manual_zero_offered says whether aborting sends that zero). A draft also shows whether the battery holds the first run's energy plus the plan's margin (state_of_charge, soc_energy_margin), which is what confirming it now checks, rather than a full or an empty battery.

An onboarding's history is now its own paginated endpoint, returning items and next_cursor newest first. limit defaults to 100 and accepts 1 to 500. It needs the onboarding.read scope, and internal tracker events are omitted for readers outside EBX.

Immediate dispatch now takes the same control payload as scheduled dispatch: control_mode, with bess_power_mw and poc_power_mw. Nothing needs to change today — set_power_mw is still accepted and maps to bess_power_mw, and control_mode defaults to bess_only. It is deprecated and no longer required, so send bess_power_mw instead when convenient. Sending set_power_mw together with bess_power_mw or power_mw is rejected.

The dispatch these return now carries control_mode, bess_power_mw, poc_power_mw and pv_power_mw alongside the deprecated set_power_mw. control_mode is one of passive, bess_only, bess_and_poc or poc_only; bess_and_poc and poc_only apply to co-located assets, and pv_power_mw is reserved and always null.

Assets carry an ancillary-service ramp pair, as_max_pos_ramp_rate_mw_per_sec and as_min_neg_ramp_rate_mw_per_sec, in MW/s like the wholesale pair. Both are optional and read null until an asset is backfilled; treat null as "use the wholesale ramp".