DIA Renewal: Reconcile Every SLA Credit Before You Accept the New Rate

A dedicated internet renewal can look great on paper while the service history tells a different story.

The proposal shows the new rate, bandwidth, and term. The carrier report shows contract uptime. Your network team remembers outages, slow recovery, and tickets that went nowhere. Finance cannot find the credits. Nobody can explain which record controls.

Do not negotiate another term until those records reconcile by circuit.

The point is not to chase a few dollars in credits. The point is to decide whether the provider delivered the reliability, support, and remedies you thought you bought. That evidence should affect the next rate, the next SLA, and whether the incumbent deserves the renewal at all.

Start with the signed SLA, not the sales promise

A carrier product page may advertise an uptime guarantee. Your actual remedy depends on the agreement, order, service guide, product, access design, location, and exclusions incorporated into your deal.

AT&T currently markets Dedicated Internet with a “100% uptime guarantee” and says a customer can request a service level credit if the service goes out. The same page directs buyers to the AT&T Service Guide for the governing SLA details.

That second sentence matters more than the headline.

Lumen’s current Service Level Agreement, version July 28, 2026, shows why. It assigns qualifying services to availability tiers, uses different credit schedules, excludes some services from availability credits, defines excused outages, and limits how credits are calculated. It also says requests generally must be submitted within 60 days after the end of the month in which the event occurred unless the agreement says otherwise.

Your contract may be different. Read your documents as one package:

Master agreement and amendments
Product service attachment or guide
Every active circuit order
SLA schedule and incorporated online terms
Renewal proposal
Provider notices that changed terms
Billing dispute and credit language

Have legal or procurement confirm which document wins if the terms conflict.

Build one outage register by circuit

Do not review reliability at the billing-account level. One account may contain several circuits, access providers, locations, and SLA treatments.

Create one row for every outage or material degradation during a review period that fits your risk. Twelve months is a useful starting point when the records exist. Go longer for chronic problems or critical sites.

Field	Evidence to capture
Circuit	Provider service ID, customer circuit ID, location, bandwidth, and business role
Event	Start, detection, ticket open, provider acknowledgment, restoration, and ticket close times
Impact	Total loss or degradation, affected applications, users, transactions, and workaround
Measurement	Internal monitor, provider portal, SD-WAN, firewall, application telemetry, and site report
Cause	Provider finding, local access issue, customer equipment, power, maintenance, unknown, or disputed
SLA treatment	Availability, latency, packet delivery, repair, response, or another covered measure
Exclusion	Maintenance, customer action, force majeure, equipment boundary, high utilization, or another term
Remedy	Credit requested, approved, denied, pending, or missed
Billing proof	Credit memo, invoice line, amount, and date applied
Renewal action	Keep, correct, shorten, compare, or replace

Keep the raw evidence behind each row. An outage spreadsheet built from memory will fall apart when the carrier challenges a timestamp.

Reconcile the clocks before arguing about uptime

Your users may say the circuit failed at 9:02. Monitoring may show packet loss at 9:03. The service desk may open a ticket at 9:14. The provider may acknowledge it at 9:28 and mark service restored at 10:01. The ticket might remain open until the next morning.

Those timestamps are not interchangeable.

Lumen’s current SLA defines service unavailability for qualifying internet service as the complete inability to deliver IP packets from the customer site to the network, excluding excused outages. It measures that unavailability from the time a trouble ticket is opened until service is restored.

That definition creates a practical buyer problem. If your monitoring detects an outage twelve minutes before the ticket opens, those twelve minutes may matter to the business while sitting outside the provider’s measurement window.

Build a timeline for every material event:

First technical evidence of failure
First user or application impact
Internal alert and escalation
Provider ticket opened
Provider acknowledgment
Troubleshooting and dispatch milestones
Technical restoration
Business service restored
Ticket closed

Compare the provider’s SLA clock with the business-impact clock. One determines the contractual remedy. The other tells you whether the service met the company’s need.

Find out what the SLA actually measures

“The network met SLA” is not enough. Ask which metric, path, location, period, and system produced that answer.

Some performance commitments are network averages rather than measurements of your application path. Verizon’s current Global Latency and Packet Delivery SLA says its network latency measurements are averaged over a calendar month between designated hub routers. The page also lists a force majeure exclusion for latency credits.

Lumen’s SLA says its internet latency and packet-loss performance levels are measured between designated points of presence and averaged monthly. Its end-to-end service levels apply only to certain services with enhanced reporting, and some measurements exclude periods when contracted bandwidth utilization exceeds 70 percent.

For each commitment, record:

The exact covered service and location
Whether the measure is site to network, point of presence to point of presence, or end to end
Whether the threshold uses a monthly average or an individual event
Whose monitoring system controls the calculation
When the provider clock starts and stops
Whether degradation counts or only complete unavailability
Which maintenance, utilization, access, equipment, and customer events are excluded
What remedy applies when the threshold is missed

A backbone average can be healthy while one site’s access circuit is failing. A monthly average can hide a short incident that caused serious business damage. A dashboard can show reachability while voice or cloud applications remain unusable.

Measure the contract. Also measure the workflow the business cares about.

Audit every eligible credit through the invoice

A service credit is not real because a representative said it was approved. Trace it to the bill.

For every potentially eligible event, answer:

Did the team open the required ticket through the required channel?
Did the request identify the affected service clearly?
Was the claim submitted before the contractual deadline?
Did it include the timestamps and evidence the provider requires?
Did the carrier approve, deny, or ignore it?
If denied, which exact term supported the denial?
If approved, where did the credit appear?
Was the amount calculated against the correct recurring charge?

Lumen’s current terms provide a useful example of the details buyers miss. The agreement says credits may be unavailable when a request is late, that monthly credits cannot exceed the specified recurring charge basis for affected services, and that discounts and special pricing are deducted before credits are calculated. It also says Lumen’s maintenance log and trouble-ticketing systems are used to calculate service-level events.

That is why your claim process needs an owner. Somebody must know the deadline, preserve the evidence, submit the request, challenge unsupported denials, and confirm the billing result.

Add a monthly control instead of waiting for renewal:

Export outage and degradation events.
Match them to carrier tickets.
Test each event against the SLA.
Submit eligible claims.
Review denials against the signed language.
Confirm approved credits on the invoice.
Carry unresolved items into the vendor review.

If claims routinely expire because nobody owns them, fix the operating process. Do not sign another SLA and pretend the remedy will manage itself.

Put the dollars in perspective

Credits matter, but they rarely equal the cost of the outage.

A credit may apply only to the affected circuit’s monthly recurring charge. The provider’s financial remedy and your business impact can be nowhere close.

Estimate the impact of each material event using defensible records:

Employees or locations unable to work
Customer interactions delayed or lost
Transactions interrupted
Overtime and incident labor
Emergency purchases or cellular usage
Missed service obligations
Work required to restore applications after connectivity returned

Use a range when the evidence is incomplete. Separate measured cost from estimated exposure.

This is the decision you need: does the circuit’s price, reliability, support, and continuity design make sense for the workflow it carries?

A strong SLA on one circuit does not replace tested redundancy. If the site cannot tolerate the outage window, use the internet circuit diversity buying guide and test the backup path. If your records do not even agree on which circuit is installed where, complete the ISP service ID reconciliation first.

Turn the history into renewal terms

Once the outage and credit records reconcile, give every circuit one decision.

Renew as proposed when the service met the required performance, support worked, remedies were handled correctly, pricing is competitive, and the contract still fits the site’s risk.

Renew with corrections when the provider fits but the SLA scope, measurement boundary, claim process, escalation path, reporting, or billing treatment needs written changes.

Use a shorter bridge when the renewal deadline is close and material outages, credits, or access-design questions remain unresolved.

Run a competitive review when recurring failures, weak support, poor transparency, commercial risk, or a better local option justifies testing the market.

Replace or redesign when the current provider or single-circuit architecture cannot meet the business requirement at an acceptable cost.

Put the required corrections into the signed record. Ask for:

A circuit schedule that maps each service ID to its SLA tier
Clear availability and performance definitions
The measurement source and reporting access
Claim steps, deadlines, and required evidence
Named escalation paths for critical incidents and billing disputes
Credit calculation examples based on the actual service charges
Chronic outage or termination rights, where available
Transition and continuity terms if service must be replaced

Do not accept a better rate as payment for an unresolved service history. A discount can reduce the price of the next outage. It does not fix the outage, the escalation process, or the contract.

If your DIA agreement is approaching renewal, schedule a technology vendor negotiation review. Bring the agreement, SLA, circuit inventory, outage tickets, monitoring records, claim notices, credit memos, invoices, and renewal proposal. Catch Advisors can help you reconcile the evidence and decide whether to renew, correct, bridge, compare, or replace before you sign.