Reducing unnecessary ZREC notifications

The Concentrator reports problems on its own receiver lines as ordinary signals on generated ZREC accounts: ZREC + the two-character concentrator ID + the two-digit port number, so ZREC0402 is concentrator 04, port 02. Each of these reaches the Alarm Queue like a customer alarm. A site with paired concentrators, several Front Ends and short line timeouts can produce a steady stream of them for conditions that resolve on their own. Four settings decide how many reach an operator: the line timeout on the Concentrator, the Restore Delay on the zone, how the ZREC accounts are grouped into installations, and the Class of the zones nobody needs to act on. How ZREC signals travel and what operators do with them is covered in the Signal Path module.

How the module progresses

1. Set each receiver line's timeout to match how often that receiver really talks.
2. Hold short line faults out of the queue with a Restore Delay.
3. Place related ZREC accounts under one installation, so a pair of concentrators or Front Ends raises one incident.
4. Set the zones nobody acts on to log only, including the signals a receiver produces when it moves from concentrator A to B.

1. Set the line timeout

Task: choose a timeout for each receiver line that catches a dead line without firing on a quiet one.

TOUT2 is the line watchdog. Each port's concport_N.txt file on the Concentrator carries TOUT2, the number of seconds the Concentrator waits for any traffic from the receiver. When it expires, the Concentrator raises zone 1 (Line timeout) on that port's ZREC account; when traffic returns, it raises the restore. Every expiry is therefore a pair of signals, and a timeout set too tight produces a pair every time the receiver goes briefly quiet.

Size it from the receiver's idle interval.

Phone lines have their own timeout. The tout value on each telephone line in the same file sets the expected gap between signals on that line. Its expiry raises zones 301–3nn (phone-line timeout, line 1–n) on the port's ZREC account. A phone line that legitimately carries little traffic needs a tout long enough to cover its quietest period.

Out-of-service ports. A port that is deliberately out of use can have timeout monitoring switched off with SuppressTOUT=1. Turn it back to 0 when the port returns to service.

Steps:

1. On the Concentrator host, open the concport_N.txt file for the port named in the ZREC account.
2. Find how often the receiver sends idle signals or is polled, from the receiver's own configuration.
3. Set TOUT2, and tout on any telephone lines, from the table above.
4. Restart the Concentrator so it reads the change, and repeat on the backup concentrator, which carries its own copy of the file.

Every option and file is described on the Concentrator page.

Example: ZREC0402 raises a line timeout and its restore several times a night, each a few seconds apart. The receiver on port 02 sends an idle signal every 30 seconds and TOUT2 is 30. Setting TOUT2=60 stops the alarms without hiding a real line failure.

2. Hold short faults out of the queue with Restore Delay

Task: let a line fault that clears within a minute or two go to the Alarm Log without reaching an operator.

Restore Delay holds the alarm while SBN waits for the restore. On a zone with Restore Required selected, the Restore Delay field holds the alarm out of the Alarm Queue. The alarm is written to the Alarm Log immediately. If the restore arrives within the delay, the operator never sees it. If it does not, a no-restore event (shown as No Res/Can) reaches the Alarm Queue. A repeated alarm on the zone restarts the timer.

It belongs on the faults that clear themselves. Line timeout (1), garbage on the line (2), communication fail (6), DSR lost (8) and phone-line timeouts (301–3nn) are the usual candidates. A receiver that drops for a moment during a network blip and recovers produces only Alarm Log entries; one that stays down still reaches an operator, delayed by the length of the Restore Delay. Set the delay shorter than the time the site accepts before technical staff must know about a dead line.

The field needs the extended zone fields. Restore Delay is shown in the Zone Details window only when option ba123 is on.

Steps:

1. Open the ZREC installation in #559 and choose the Zones tab.
2. Open the zone with Get One Zone [ALT+F1], then Change Record.
3. Select Restore Required, set the Restore Zone to the same zone, and enter the Restore Delay.
4. Save [ENTER].

Restore Required and the Restore fields are described in the Zones module; forcing restores into the queue is in Restore Exception.

Example: zone 1 on ZREC0402 is given Restore Required and a Restore Delay of 120 seconds. A 40-second outage writes an alarm and a restore to the Alarm Log and nothing to the queue. A 10-minute outage produces one No Res/Can alarm two minutes in.

3. Place related ZRECs under one installation

Task: set up the ZREC accounts so that one fault seen from two places raises one incident.

Paired equipment reports the same fault twice. Paired concentrators are numbered 10 apart (primary 04, backup 14), and each stamps its own ID on the alarm, so a receiver line on port 02 has two accounts: ZREC0402 and ZREC1402. Each Front End instance likewise writes its startup signal on its own account. With one installation per account, a fault seen by both members of a pair arrives as two separate alarms, each with its own action plan.

One installation, several CIDs. An installation can carry more than one CID in its CID grid. Put the accounts that describe the same equipment on one installation:

Shared zones and one Group make it one incident. Create the zones once with CID %, and tick Shared Zone Pickup on each CID in its CID Zoom window, so the same zone answers a signal from either concentrator or Front End. Give those zones the same monitoring Group: further signals in the group join the open incident instead of raising a new alarm line. The operator works one incident and its history shows which member of the pair reported what.

Steps:

1. In #559, create one installation per receiver line (or per Front End pair), named for what it supervises.
2. On its CID grid, add each ZREC account from the pair as a CID, and tick Shared Zone Pickup on each.
3. On the Zones tab, create the zones with CID %, one Group code for the zones that belong together, and the technical escalation action plan.
4. If the accounts already exist as separate installations, set their monitoring status to inactive once the combined installation is receiving their signals.

Shared zones, Shared Zone Pickup and Group are described in the Zones module.

Example: a network fault cuts the receiver on port 02 off from both concentrators. Before: ZREC0402 and ZREC1402 each raise a line timeout, and two operators may pick up the same fault. After: both CIDs sit on CONC 04/14 PORT 02, the second timeout joins the first incident, and one operator handles it.

4. Set the zones nobody acts on to log only

Task: keep a record of expected, harmless events without sending them to an operator.

Class decides where a zone's signals go. The Class field on a zone takes one of three values:

Use = for events that are expected and need no action. Candidates are the Front End startup signal (zone 1 on a Front End account) when restarts are planned routine, and receiver self-test (zone 10). Use = in preference to -: the Alarm Log keeps the history needed when a pattern later points to a real fault.

A receiver moving from concentrator A to B

The Concentrator raises no signal of its own for a move. Each concentrator watches only its own lines, so a receiver moving from CONC A to CONC B reaches SBN as ordinary line signals on two accounts:

The restore on B arrives first and the timeout on A follows TOUT2 seconds later, so the alarm on A is the signal that reaches an operator. A move back from B to A produces the same signals with the accounts swapped.

Log only means Class = on zone 1, and zone 8 on serial lines, for that port on both concentrators. The same zone 1 alarm is what a receiver that has failed altogether produces, so with this setting a dead receiver line also goes only to the Alarm Log. Agree with the site how a dead receiver will be noticed before making the change, for example through the Test Interval on the customer installations on that line, which raises a missing-test alarm when their panels stop reporting.

Steps:

1. In the Alarm Log, search the ZREC accounts for that port around the time of a known move and confirm the alarm is zone 1 (and zone 8 on a serial line).
2. Open the ZREC installation in #559, choose the Zones tab, and open zone 1 with Get One Zone [ALT+F1], then Change Record.
3. Set Class to = and Save [ENTER]. Repeat for zone 8 on a serial line.
4. Repeat on the ZREC installation for the same port on the other concentrator. When both accounts sit on one installation with shared zones (section 3), one change covers both.

Example: every failover of the receiver on port 02 from CONC A to CONC B puts a line timeout on ZREC0402 in the queue, and operators close it without action. Setting Class = on zone 1 of CONC 04/14 PORT 02 keeps every failover in the Alarm Log and out of the queue. The site relies on the Test Interval of the installations on that receiver to report a receiver that stops altogether.

Summary