Event record

Read one downtime event's facts, alarms, classification and journal, and change it with a recorded reason.

Events is early access. Changing an event needs permission to write events, which operators and admins have.

Select an asset name in the Events list to open that event's record. The record has a stable address, so the list, a contract's availability result and a colleague's link all open the same page. Events at the top returns to the list.

An event record: its state and classification badges, the change buttons, and the Record facts

The heading names the asset, then the site and how the event began. The badges show the state, Classified or Unclassified, and, when an agent has proposed a classification nobody has answered yet, for example 1 agent proposal to review. The event ID follows.

Record

Record lists Started, Ended, Duration, State, Origin, Site, Asset and Recorded by, which is a person or System. An open event shows No end recorded and a duration that ends with so far. An unwatched event also shows its Unwatched reason. A deleted event shows when, and by whom, it was deleted.

The origin says how the event began:

Origin	Meaning
Recorded by a person	Someone recorded the downtime
Opened by an alarm	An alarm breach opened it. Opening alarm links to that alarm
Opened by a maintenance plan	A maintenance plan opened it
Imported	It came from an import
Computed by the state engine	The state engine opened it

Linked alarms

Linked alarms lists each breach linked to the event, with its Breach time and its Recovery. A recovery is only what the alarm measured. An administrative close never counts as one, and the row shows its reason separately. Each row says whether the alarm Opened this event or Joined after it opened, and Open alarm opens it. When the event closed while a breach was still active, the row says No recovery recorded. The breach was still active when this event closed.

Change an event

Operators and admins see the changes the event's state and origin allow. Each opens a dialog that says what the change will do before you send it.

Choose Close event, Trim, Split, Merge or Delete.
Fill in the times or the choice the dialog asks for. Times are in UTC, to the second.
Enter a Reason (required), up to 2,000 characters. It is kept in the event's journal with your name.
Choose the dialog's action button, for example Trim event. Cancel closes it without a change.

The Trim this event dialog with Start (UTC), End (UTC) and the required reason

Change	What it does
Close event	Sets the end of an open or unwatched event. The end must be after the start and not in the future. Lost output is then computed for the closed range
Trim	Moves the start later, the end earlier, or both. A trim never widens an event. An open event has no end to trim; close it first
Split	Ends this event at the split point, which must lie strictly inside it. A new event takes the rest, with the same origin, no classification and no lost output
Merge	Joins this event with the event directly before or after it on the same asset. The earlier event survives and takes the later one's end and state; the later one is deleted and keeps its own journal, classification and lost output
Delete	Removes the event from every list and every availability figure. Its record, classification history and journal stay readable at its address

Some changes depend on how the event began, because a system-opened event's boundaries were measured. Only events recorded by a person can be split, and only an event recorded by a person can be absorbed as the later event of a merge. When a change is not offered, a line under the buttons says why, for example Close: Already closed. Trim moves its end earlier. A deleted event takes no further changes.

After a change, a line above the record says what happened and links any new or surviving event, for example Open the new event after a split. The record and every Events list then refresh.

A trim, split or merge leaves the stored lost output on the old range until it is computed again. See Lost output.

Journal

Journal lists every change to the event, oldest first. Entries are never edited or removed. Each entry shows when it happened, what happened, for example Opened, Closed, Trimmed, Classified or Lost output computed, any change of state, and who acted: a person, System, Agent or Integration. A person's reason shows as Reason:, and an entry that touched another event links it, for example New event after a split.

The Lost output section and the Journal with the opening, closing, lost-output and classification entries

When a change does not go through

The dialog keeps what you typed and shows the reason. The record re-reads, so you see the state that won.

Someone changed this event first. Check the refreshed record, then submit again.
That interval would overlap another event of this asset. The dialog links the event in the way.
The database is busy. The change may or may not have saved. Read the journal before you submit again. The submit button stays paused for the few seconds the message names, while the database recovers.
No usable answer came back. The change may have saved. Read the journal before you submit again.

When the record itself cannot be read because the database is busy, the page says The event database is busy. and Nothing is wrong with this event. Retry becomes available once the countdown ends.

An address that does not match an event you can read shows Event not found, whether the event never existed or belongs to someone else.

Classification

Classify this event, or review an agent's proposal.

Lost output

Read this event's lost output and compute it again.