A format specification for NorthAI's first customer SLA. Uptime tiers, response-time matrix, escalation paths, credit schedule, and FedRAMP security incident carve-out. Populated once rate card and customer details are confirmed.
This artifact is a format specification, not a filled-in contract. The SLA structure below reflects FedRAMP-aware defaults grounded in industry practice for federal SaaS vendors. The content fields marked [CONFIRM WITH TIM AND STEPHANIE] require NorthAI's internal decisions on rate cards, customer deployment environments, and escalation personnel before this template becomes a customer-facing document.
The engagement produces the filled version. This format shows what the engagement delivers. No section below constitutes a commitment by HARBOR Initiative LLC or NorthAI to any service level, pricing term, or legal obligation.
Federal customers expect uptime commitments expressed as annual availability percentages with corresponding allowed downtime windows. Two tiers are the standard structure for a federal SaaS product at the Level 2 productized service stage.
| Tier | Availability target | Allowed downtime (monthly) | Allowed downtime (annual) | Applicable to | Status |
|---|---|---|---|---|---|
| Standard | 99.9% | 43.8 minutes | 8.76 hours | Base subscription tier (CLIN 0001) | [CONFIRM: Is 99.9% the correct standard floor?] |
| Premium | 99.95% | 21.9 minutes | 4.38 hours | Agencies requiring higher availability SLA (optional upgrade, CLIN 0001 variant or CLIN 0003 T&M) | [CONFIRM: Is 99.95% the right premium ceiling given current infrastructure?] |
Measurement window: Calendar month. Availability is calculated as: ((Total minutes in month - Downtime minutes) / Total minutes in month) x 100.
Excluded from downtime calculation (standard exclusions):
Monitoring: Uptime tracked via [CONFIRM: monitoring tool, e.g., StatusPage, Datadog, AWS CloudWatch]. Status page URL: [CONFIRM URL]. Customers have read access to status dashboard.
Four severity levels map to the standard federal IT incident classification framework. Response times below are defaults grounded in common federal SaaS SLAs. Adjust based on NorthAI's current support team capacity and customer tier.
| Severity | Definition | Examples for NorthAI | Initial response (Standard) | Initial response (Premium) | Target resolution | Escalation trigger |
|---|---|---|---|---|---|---|
| Sev 1 · Critical | Complete service unavailability. Production environment down. No workaround available. Federal mission impact. | Analytics platform unreachable; authentication service down; data ingestion pipeline halted with no bypass | Within 1 hour (24x7) | Within 30 minutes (24x7) | 4 hours (restore service); 24 hours (root cause) | Auto-escalate to Engineering Manager at 1 hour if unresolved. VP Engineering at 2 hours. |
| Sev 2 · High | Major feature degradation. Core functionality impaired but workaround exists. Limited mission impact. | Report generation latency exceeding 5x normal; specific data source connector failing; dashboard rendering incomplete for one or more agencies | Within 4 hours (business hours: 08:00-20:00 ET) | Within 2 hours (business hours) | 24 hours (workaround confirmed); 72 hours (resolution) | Auto-escalate to Engineering Manager at 4 hours if no workaround confirmed. |
| Sev 3 · Medium | Minor feature degradation or non-critical service disruption. Workaround available. No mission impact. | Non-critical report export failing; UI rendering issue in specific browser; integration with optional data source intermittent | Within 1 business day | Within 1 business day | 5 business days | Escalate to TAM at 2 business days if no update provided. |
| Sev 4 · Low | Cosmetic issue, documentation request, general inquiry, or enhancement request. No service impact. | UI label inconsistency; documentation update needed; feature request submission; general configuration question | Within 3 business days | Within 2 business days | Next scheduled release or 30 business days (as appropriate) | No escalation path; tracked in product backlog. |
[CONFIRM WITH TIM AND STEPHANIE: Current support team coverage hours and on-call rotation capacity. 24x7 Sev 1 coverage requires defined on-call rotation. Confirm whether NorthAI can commit to 24x7 response before offering the Standard SLA Sev 1 commitment above.]
Three escalation tiers. Personnel names are placeholders; populate with actual NorthAI team members before customer delivery.
| Tier | Role | Name | Contact | Escalation trigger |
|---|---|---|---|---|
| Tier 1 | Technical Account Manager (TAM) | [CONFIRM NAME] | [CONFIRM EMAIL AND PHONE] | First point of contact for all Sev 1-4 issues. Owns ticket through resolution. Escalates to Tier 2 at defined thresholds. |
| Tier 2 | Engineering Manager | [CONFIRM NAME] | [CONFIRM EMAIL AND PHONE] | Sev 1 unresolved at 1 hour; Sev 2 unresolved at 4 hours (no workaround). Owns engineering response and customer communication during active Sev 1/2 incidents. |
| Tier 3 | VP Engineering (or CTO equivalent) | [CONFIRM NAME, Tim Otto?] | [CONFIRM EMAIL AND PHONE] | Sev 1 unresolved at 2 hours; any security incident triggering FedRAMP notification window; customer-escalated executive request. Final technical authority on incident response decisions. |
Federal customers may also escalate through the Contracting Officer (CO) or Contracting Officer's Representative (COR) named in the task order. NorthAI's escalation path maps to the COR-side as follows:
Service credits are the standard remedy for SLA misses under a federal SaaS contract. Credits apply to the billing period in which the SLA miss occurred. Credits do not cascade to subsequent periods. Credits are the sole remedy for uptime SLA misses; they do not waive or reduce other contractual rights.
| Monthly availability (Standard tier) | Monthly availability (Premium tier) | Credit applied to next invoice |
|---|---|---|
| 99.9% or above | 99.95% or above | 0% (SLA met; no credit) |
| 99.0% to 99.89% | 99.5% to 99.94% | 10% of monthly subscription fee |
| 95.0% to 98.99% | 95.0% to 99.49% | 25% of monthly subscription fee |
| 90.0% to 94.99% | 90.0% to 94.99% | 40% of monthly subscription fee |
| Below 90.0% | Below 90.0% | 50% of monthly subscription fee (maximum credit) |
[CONFIRM WITH TIM AND STEPHANIE: Is the 50% annual cap acceptable? Does legal counsel need to review the credit structure before customer delivery?]
Security incidents that trigger FedRAMP notification requirements are subject to a separate timeline and procedure that supersedes the standard SLA response-time matrix in Section 2. This carve-out is not optional for a FedRAMP-authorized product; it reflects the mandatory federal notification framework.
| Timeline | Required action | Recipient |
|---|---|---|
| 1 hour | Verbal or written notification of confirmed or suspected security incident to customer COR and CO | Customer Contracting Officer (CO) and Contracting Officer's Representative (COR) |
| 2 hours | Initial incident report (preliminary scope, affected systems, containment steps in progress) | CO, COR, NorthAI compliance lead, NorthAI VP Engineering |
| 72 hours (mandatory) | Formal federal incident notification submitted to FedRAMP Program Management Office, US-CERT, and all customer agencies using the authorized system | FedRAMP PMO, US-CERT, all customer agencies (agency ISSOs) |
| 7 days | Incident summary report: confirmed scope, root cause (if determined), remediation steps taken, current status, POA&M entry for residual risk | All customers, FedRAMP PMO, 3PAO (if assessment update triggered) |
| 30 days | Root cause analysis and final incident report. Updated SSP if boundary was affected. 3PAO notification if significant boundary change occurred. | FedRAMP PMO, all customers, 3PAO, CO/COR |
Critical note: The 72-hour FedRAMP notification window is a regulatory requirement, not an SLA commitment. Failure to notify within 72 hours can result in suspension or revocation of FedRAMP authorization. Standard SLA credits (Section 4) do not apply during a declared security incident. A separate incident remediation process governs customer remedies during and after a security event.
Routine technical issues (Sev 1 outages, performance degradation, failed deployments) that do not involve unauthorized data access or boundary compromise are governed by Sections 2-4, not this security incident carve-out.