Stop paying for a backup link that shouldn’t be carrying traffic.
The most expensive WAN problems have no symptoms. Service quality is fine, so nobody looks. A metered backup stuck in service, or a circuit running at ten percent, shows up only on the invoice.
Recognize it
You’re here if
- A metered 4G or 5G link has overage charges and the primary is healthy.
- A premium circuit has been under 15% utilization for weeks.
- Cost review is next month and nobody knows where the spend went.
Do this
5 steps, in order
Find a healthy primary with a busy metered backup
Pick a site. Compare the primary’s Utilization with the backup’s over the widest window.
Open Explorer- Look for
- The primary near idle and available, the metered backup carrying the traffic.
- What it means
- That is a stuck failover. Service is fine and the meter is running.
Clear the hold-down
Review the recommendation to reset the preferred-path policy so the site returns to the primary.
Open Remediation- Look for
- A reversible action, gated for approval.
- What it means
- Confirm afterwards that the backup falls idle and overage stops.
Add a detector so it can’t recur silently
In KPI Studio, define a link KPI scoped to the 4G and 5G link types: sustained average utilization. Test it on history, then operationalize it.
Open KPI Studio- Look for
- A backtest showing when it would have fired on the stuck-failover weeks.
- What it means
- A metered link that carries traffic for days should tell you, rather than the invoice.
Find idle capacity
Look for primaries running at low utilization while meeting their targets.
Open SLOs- Look for
- Circuits at 8 to 15% utilization at peak.
- What it means
- Right-size the committed rate, or move baseline traffic to a cheaper path and keep the premium circuit for latency-sensitive flows.
Bank the number
Add the projected monthly saving to a board so it is visible at the cost review.
Open Boards- Look for
- A saving someone owns.
- What it means
- An unowned saving is not a saving.
Then decide
What you found decides what you do
IfBackup is metered, primary is healthy and idle
ThenStuck on backup. Reset failback and add the detector.
RB-COST-STUCK-BACKUPIfPrimary is consistently under about 15% at peak
ThenRight-size or re-route baseline traffic.
RB-COST-UNDERUTILIfOverage coincides with a bulk transfer on a metered link
ThenMove the transfer to the primary or schedule it off-peak. Voice dies at the same hour daily →
Watch out
Where this goes wrong
- Do not right-size on one quiet month. Look at the widest window and at seasonality before you cut capacity.
The stuck backup is cleared, a detector exists, and the idle capacity has an owner and a number.
Keep going
Often next
Claim an SLA credit
Find the breaches, assemble dated evidence, and build the scorecard for the renewal.
~20 minFinance and vendor management, Network lead, Service ownerSomething is downA circuit went down
Confirm traffic actually moved, size the damage on backup, and set the failback condition.
~10 minNOC engineer, Network leadRun the NOCBuild your NOC wall
Choose the questions the wall answers, build a board for them, and put it on a screen.
~25 minNOC engineer, Service owner, Network lead