Remote hands SLA: the 5 clauses your provider does not want in writing (and why you should push for them)

The remote hands SLA is the most important document in the contract with your field services provider. And it is also the one your provider will want to leave as ambiguous as possible. The specific clauses on response time, time on site, escalation, retry, and penalties are what separate a service that works from one that leaves you alone at 3 AM when your network goes down. This article breaks down the 5 clauses the provider would rather leave vague, what each one has to say, and how to push to get them in writing.

What follows applies to data centers with contracted remote hands services (colocation, site provider, or a provider specialized in smart hands). If your operation is 100% remote with no physical presence, some clauses apply differently. The suggested time ranges are for production data centers with an equivalent Tier III SLA. For Tier IV, the times are stricter.

Clause 1: response time from report to acknowledgment

This clause defines how long the provider takes to confirm that it received your ticket and is working on it. It is NOT the time it takes to reach the site. It is only the time it takes to say “we hear you”. Why the provider doesn’t want it specific:

1. If it says “we confirm within a maximum of 15 minutes”, the provider has to have 24/7 on-call staff dedicated to reading tickets. That costs money they don’t want to take on.
2. If it says ‘we confirm within a maximum of 5 minutes’, you need an automated system with escalation. The provider prefers vagueness so that the real acknowledgment takes 30-60 minutes and there is no penalty.

What the clause should say: “The provider confirms receipt of the incident by email, phone call, or message to the client system within the first 15 minutes for emergencies reported during operational hours, and within 30 minutes for emergencies reported outside operational hours. The confirmation includes: incident ID, the operator person who received the report, and estimated time to attention.”

Suggested penalty if breached: a 5% discount on the monthly contract cost for each incident where the acknowledgment time is exceeded without documented justification.

Clause 2: on-site response time (hands-on-site)

This is the critical clause. It defines how long the provider takes to have a remote hand physically in front of your rack after your report. The typical ranges you should ask for:

1. Tier I (critical incident, production down): a maximum of 2 hours from acknowledgment to physical presence. For Tier III data centers with redundancy, this is realistic in metropolitan areas. For Tier IV or remote DCs, it should be 1 hour.
2. Tier II (urgent, non-critical but time-sensitive incident): a maximum of 4 business hours. This covers tickets like ‘replace this disk before the rest fill up’ or ‘label the ports before the new equipment arrives’.
3. Tier III (scheduled ticket, non-emergency): within 24-48 business hours, with a scheduling window confirmed at least 4 hours in advance.

Why the provider wants it vague: if it says ‘a maximum of 2 hours’, they have to have on-call technicians in areas close to ALL the data centers they serve. If the provider serves 20 DCs in 5 cities, it needs at least 1 on-call technician per city. That is a fixed staffing cost many don’t want to absorb.

How to push: define the SLA per tier and ask the provider to quote each tier separately. If the Tier I price is prohibitive, negotiate Tier I only for 2-3 critical DCs and standard Tier II for the rest. What matters is that it is in writing which DCs have which tier.

Clause 3: escalation through on-site work when there is a blocker

This clause covers the case where the remote hand arrives on site but cannot resolve it because it needs a decision from the manufacturer, the upstream provider, or your own team. Why the provider wants it vague: if it is clear, the provider is obliged to manage the escalation, and that means coordination staff it does not have today.

What the clause should say: “In the event that the remote hand cannot resolve the incident because it requires intervention from the manufacturer, the upstream provider, or authorized client personnel, the provider manages the communication with the third party, documents the blocker, and maintains physical presence until resolution or until the rollback order is lifted. The escalation management time is the provider’s responsibility and must be reported in the ticket.”

Typical case: your UPS fails and the remote hand arrives but needs the manufacturer (Vertiv, Schneider, etc.) to tell it what to do. Without an escalation clause, the remote hand leaves and you end up waiting for someone else to show up. With the clause, the provider coordinates the call with the manufacturer, keeps the remote hand on site, and you only handle the final technical decision.

Clause 4: right to retry at no charge in the case of a remote hands failure

This is the clause that covers the worst case: the remote hand performs a task and does it wrong, requiring a second intervention. Why the provider wants it vague: if it clearly says ‘retry at no cost’, it knows it will have to accept that its technicians make mistakes and redo the work without charging you extra.

What the clause should say: ‘In the event that the remote hands intervention results in a failure due to an operational error by the technician (wrong slot identification, wrong cable connection, incorrect configuration, physical damage to client equipment), the provider performs the correction at no additional cost to the client, including any spare part or replacement of the component damaged by the intervention. The provider’s liability covers the incident and its operational consequences within the 72 hours following the original intervention.’

Why it matters: if the remote hand plugs a network cable into the wrong port and causes an outage, the cost is not just the corrective intervention but the service unavailability. Without a no-cost retry clause, the provider is going to want to charge you for the second visit and you have no recourse. With the clause, it is clear that the remote hand is responsible for its error until full resolution.

Clause 5: measurable penalty for SLA breach

The clause that closes everything. Without a penalty, the other 4 are suggestions. Why the provider wants it vague or null: because it knows that any documented SLA without a penalty is, in practice, optional.

What the clause should say: ‘For each incident where the provider breaches the acknowledgment time, the on-site time, or the escalation management time defined in clauses 1-3, the provider applies a 10% discount on the monthly contract cost. The discount accumulates and is applied on the following month’s invoice. The provider keeps a monthly record of incidents with actual vs committed times, signed by both parties.’

Acceptable variants: (a) monthly accumulated discount (10% per incident, no cap), (b) capped discount (maximum 30% of the month), (c) service hour credit (1 free hour per breached incident), (d) right to terminate the contract without penalty for sustained breach (3 or more serious incidents in 6 months). Pick the one that works best for you depending on the maturity of the relationship with the provider.

How to get these clauses into a current contract

If you already have a contract with your provider and it does not have these 5 clauses, adding them is not trivial. The realistic path:

1. Request an SLA review meeting. Frame it as «operational alignment», not as “I want to complain”. The provider gets defensive if you frame it as a conflict.
2. Bring concrete data: 3-5 recent incidents where the provider missed the times. Without data, your request is abstract.
3. Propose the 5 clauses as an addendum to the existing contract. An addendum gets signed faster than renegotiating the whole contract and sets a precedent for the next renewal.
4. If the provider refuses any penalty clause, that is information: your provider knows it will not comply and prefers to stay in the ambiguity. Weigh that answer when evaluating the relationship.

If your current provider rejects the 5 clauses but complies with documented response times and tickets without serious incidents in the last 6 months, you can accept a looser SLA. But document it in writing: “We operate with an informal SLA, quarterly metrics reviewed in a meeting, no formal penalty”. At least it stands as a precedent if quality degrades.

How to choose a provider if you are starting from scratch

If you do not have a remote hands provider, the 5 clauses serve as an objective filter for proposals:

1. If the provider already includes the 5 clauses in its standard contract on day 1, it is a mature provider that knows how to comply.
2. If the provider says ‘it depends, we will look at it case by case’, it is an improvised or new provider in the market. It works for simple tickets but not for critical incidents.
3. Ask the provider for historical metrics: average ticket resolution time, incidents per month, actual vs committed time over the last 6 months. If it does not have them, it is a new provider with no track record.
4. Verify that it has real geographic coverage. A provider that says ‘we serve the whole republic’ but only has technicians in CDMX is not going to meet the Tier I SLA in Monterrey.

The SLA is not a document you sign once and forget. It is the operational foundation of your relationship with your site provider. If the SLA is ambiguous, the relationship will be ambiguous. If the SLA is specific, operational problems have a clear escalation and resolution path.

Sources

1. Uptime Institute — Operational Sustainability Tier Certification (Tier III/IV SLAs) — https://datacenter.uptimeinstitute.com/
2. ICCP — International Computer Communications Service Provider association: SLA guidelines — https://www.iccp.org/
3. ISO/IEC 20000-1:2018 — Service management system requirements — https://www.iso.org/standard/70636.html
4. ITIL 4 Foundation — Service Level Management practice guide — https://www.axelos.com/certifications/itil-service-management/itil-4-foundation
5. BICSI 002-2024 — Data Center Operations: Field services and SLA definitions — https://www.bicsi.org/standards/data-center