Warranty Tracking
Back to Data Center Management · Scheduled Maintenance Calendar · Spare Parts Management · Escalation Contacts · Service Offerings
Warranty tracking is knowing what's still covered and by whom before a failure makes it urgent. The failure mode it prevents is specific: a machine goes down, somebody searches for the support contract, and discovers that coverage lapsed eleven weeks ago. The outage is now also a procurement exercise.
1. Every Asset Has Three Dates, Not One
These get conflated constantly, and they fail in different ways.
| Date | What ends | What you can still do |
|---|---|---|
| Warranty or contract end | Your entitlement to service under that agreement | Renew, buy a new contract, or go to a third-party maintainer |
| End of support | The vendor will no longer sell you support at any price | Third-party maintenance only, and firmware and security fixes stop |
| End of service life | Parts are no longer manufactured or held | Only what exists on shelves — see spare parts |
The second and third are the ones that drive planning, because they cannot be solved with money at short notice. They belong in growth forecasting as hard dates, not in a maintenance spreadsheet.
2. The Service Level Is Not What People Remember
"Four-hour support" is shorthand that hides the terms that matter when you are relying on it.
- Response is not resolution. A four-hour service level is usually a commitment to respond or to dispatch, not to have you running. Read what the contract actually promises, and plan recovery around that rather than around the marketing name.
- Coverage hours. Next business day means exactly that. Nine-to-five Monday-to-Friday cover on a system that matters at the weekend is a gap you should know about before the weekend.
- What is covered. Parts only, parts and labour, or onsite. Whether software and firmware support is included, which is frequently a separate line.
- Where it is covered. Onsite response is priced by location. Equipment moved to a different site, and particularly to a different country, may not carry its entitlement with it — this surprises people after a migration.
3. Entitlement Is Tied to the Serial, Which Is Why the Register Matters
Support is looked up by the vendor's serial or service tag, not by your hostname or asset tag. So warranty tracking is only as good as the asset register underneath it, and the register has to hold the vendor's identifier alongside yours.
- Record the serial, the contract or entitlement number, and the vendor's own support identifier for the agreement.
- Note that a board replaced under warranty can change the serial, which then silently detaches the machine from its own contract.
- Second-hand equipment frequently carries no transferable entitlement at all, whatever the remaining term appears to be.
- Keep this where the incident responder will find it at 03:00 — the same place as escalation contacts, and reachable when systems are down.
4. Renewal Is a Decision With a Deadline
Coverage does not renew itself, and lapsing is not a neutral state you can reverse at leisure. Vendors commonly charge a reinstatement fee, require an inspection, or simply decline to cover equipment that went uncovered.
So the useful artefact is not a list of expiry dates. It is a decision window, as in the illustration above: anything expiring in the next ninety days needs an answer now, and the answer is one of four:
- Renew with the vendor.
- Move to a third-party maintainer, which is usually cheaper and often the only option after end of support. Verify their parts access for your specific models.
- Self-insure with spares, which for commodity hardware is frequently the rational choice — see spare parts management.
- Accept the risk deliberately for equipment near replacement, recorded as a decision rather than arrived at by inattention.
All four are defensible. Only the fifth, which is not noticing, is not.
5. Coverage Should Match Importance, Not Purchase Date
Support is usually bought with the hardware and then never revisited, so cover ends up matching what things cost rather than what they now do. Review it against current criticality:
- A machine that has quietly become load-bearing may be on next-business-day cover.
- A development box bought in the same order may be on four-hour onsite.
- Equipment with no single point of failure may need less cover than its price suggests, because the recovery plan already handles its loss.
How We Approach It
- Reconcile the estate against the vendor's entitlement records, by serial. The gap between what you believe is covered and what the vendor says is covered is the first finding, and it is rarely zero.
- Record all three dates per asset, plus the actual service level terms rather than the shorthand.
- Produce the decision window: expired, expiring within ninety days, and approaching end of support.
- Match cover to current criticality, which usually means moving some cover rather than buying more.
- Put the contract and entitlement numbers where an incident responder will find them, offline.
- Set the review cadence and feed the end-of-support dates into refresh planning.
What You Get
- A coverage position reconciled against the vendor's own records, not against your purchase orders.
- Warranty end, end of support and end of service life recorded separately per asset.
- The real service level terms — response versus resolution, hours, location, what is included.
- A ninety-day decision list with a recommendation against each: renew, third party, spares, or accept.
- Entitlement details available to whoever is handling the incident, without needing a working system to look them up.
The measure is simple. When something fails at two in the morning, how long does it take to find out whether it is covered — and is the answer already written down?