Trust & security
What we attest to, what eStruxture attests to for its facilities, and what we don't claim.
A procurement reviewer should be able to read this page and know exactly where Stack Harbor's responsibility starts and ends. So it says, precisely: the controls we operate on your environment, the attestations eStruxture publishes for its facilities underneath, the evidence we provide when your auditor or client asks — and the certifications we do not hold.
Compliance position
Whose attestation covers what.
"SOC 2-aligned" is the phrase vendors use when they want the credibility of an attestation without naming whose it is. We would rather be exact.
Stack Harbor holds no SOC 2 report of its own.
We do not claim one, and nothing on this site should be read as implying one. The same is true of ISO 27001, PCI DSS, and HIPAA: we hold no certification under any of them.
eStruxture attests to its own facilities.
Our Montréal data centres include eStruxture facilities, and eStruxture states that all its facilities are ISO/IEC 27001:2022 certified and externally audited for a SOC 2 Type 2 report (eStruxture security and compliance). Those cover the physical layer — power, cooling, physical access — not the operational controls we run on top, the virtualization platform included. When you need it, we point you to eStruxture's attestation rather than restating it as ours.
We are an infrastructure operator, not an auditor.
When a client pursues SOC 2, an ISO certification, or a privacy-law determination, their auditor or privacy officer makes the attestation. We operate the infrastructure controls in scope and produce the operational evidence those determinations rest on. We do not certify a platform compliant or a site accessible — the people whose job that is, do.
Cardholder data and health records are your determination to make.
Stack Harbor does not record card numbers or CVVs in its own systems; payment data goes directly to the processor. If your workload handles cardholder data or personal health information, the scoping decision belongs to your QSA or privacy officer, and we build to the controls they specify.
Controls we operate
What is actually in place on an environment we run.
These are the controls Stack Harbor operates on environments under a Managed Cloud Ops or Managed Operations engagement. Exact scope is set per environment in the agreement — this is what an engagement is built from, not a promise that every item applies to every plan.
Backups and recovery
- 3-2-1 backup posture with off-site retention and off-site replication
- Backups included; ask us for the schedule
- Point-in-time recovery for databases
Patching and hardening
- Staged OS, control-panel, runtime, and PHP-version patching with rollback windows
- Hardening to CIS-aligned baselines
- Vulnerability scanning and vulnerability response
- Key rotation on a cadence, with an audit trail
Network and edge
- TLS on every endpoint, WAF, and intrusion detection
- Private networking and load balancing between tiers
- DDoS posture at the edge through Cloudflare and edge caching — described as a posture on purpose; volumetric attacks are an exclusion in the SLA, and we would rather you read that here than in the fine print
Access, logging, and evidence
- Account-level access controls with audit logs retained for review
- On our application platform, every SSH gate session is tied to an account, timestamped, and recorded with its target node
- Change and access records kept in a form an auditor can sample, not reconstructed after the fact
Incident response
- A 24/7/365 on-call rotation working to runbooks designed by senior engineers
- Severity 1 acknowledged within 15 minutes and Severity 2 within 60, around the clock, on plans that include incident coverage
- Triage, mitigation, communication to your team, vendor escalation, and a written postmortem on every Sev-1
- Senior engineers engaged on Sev-1 per the escalation matrix agreed up front — not paged for routine alerts
Data residency
Where the data is, written down.
Two data centres in Montréal. A workload placed there stays in Canada, and the placement is documented. Where a design intentionally uses a US or partner region, that is a decision made with you and recorded — not a default.
One distinction matters for Quebec organizations, and most hosting pages skip it: Quebec's privacy law treats communication of personal information outside Quebec — including to Ontario — as something that has to be assessed. If your posture requires data to remain in Quebec, where standby and backups live is a placement decision we make with you at design time and write down.
Headquartered in Montréal · low-latency peering
Built where your data should be.
Stack Harbor runs from two data centres in Montréal, with direct, low-latency peering to major North American and Western European markets. Workloads placed there stay in Canada; performance reaches every customer.
Service level agreement
The SLA, in numbers.
These are the figures in the agreement, not a marketing restatement of them. The acknowledged-response times are the committed times; resolution is a target. Credits are the sole and exclusive remedy for unavailability, and the SLA says so.
Technical support runs through the support desk, which is the system of record for response timing. Telephone contact reaches your account manager and is not a technical support channel.
Read the full SLA- Network and IaaS uptime
- 99.9% availability, monthly
- Severity 1 — acknowledged response
- Within 15 minutes, 24 / 7 / 365
- Severity 2 — acknowledged response
- Within 60 minutes, 24 / 7 / 365
- Severity 3 — acknowledged response
- Within 8 business hours
- Incident resolution
- Target under 24 hours — a target, not a commitment
- Service credit
- 2% per hour of downtime beyond the SLA, up to the month's fees
- General support and management requests
- 9 AM – 5 PM (EST), Monday to Friday
Partnerships and technologies
The one partnership we name, and the technologies we use.
cPanel NOC partner, through WebPros
Stack Harbor is a cPanel NOC partner — the WebPros program through which we obtain and issue cPanel & WHM licences directly. It is the reason a cPanel escalation from us goes to the vendor's engineers rather than to a forum thread, and it is the only commercial partnership this page names.
Everything else on the strip is a technology, not a partner.
Cloudflare, CloudLinux, LiteSpeed, and the other names shown across the site are technologies in use, or ones we are compatible with. We label them that way and we do not describe a technology we use as a partnership. Trademarks belong to their owners.
Evidence
Three example scenarios with a real compliance bar.
Illustrative, not specific clients. Each one shows who makes the determination, and in every case it is not us.
SOC 2 Type II readiness
A 40-person fintech SaaS and 14 auditor-flagged infrastructure gaps
Closed in twelve weeks; the Type II observation window passed; the controls kept operating under Managed Cloud Ops. The client holds the report. Their auditor made every determination.
Data residency
Customer-keyed residency for a Canadian SaaS with a US enterprise base
A 14-week dual-region rebuild with audit trails that held up under PIPEDA, Law 25, and HIPAA business-associate scrutiny. Residency became a named contract clause in four enterprise deals.
Public sector
Bilingual public-sector WordPress under a procurement framework
Residency in writing, a WCAG 2.1 AA re-test against the original conformance baseline, and procurement-aligned quarterly reporting. The body's accessibility office certifies the site; we made it possible to.
Depth reading in the knowledge base: compliance posture on cPanel — SOC 2, PIPEDA, and Law 25 and Quebec data residency in practice.
In plain words
What we do not claim.
A trust page that only lists strengths is a sales page. These are the lines we hold ourselves to.
- We hold no SOC 2 report, and no ISO 27001, PCI DSS, or HIPAA certification, of our own.
- We are not an auditor. We do not certify a platform compliant or a site accessible; we produce the evidence the people who do rely on.
- We do not guarantee uptime. We commit to 99.9% monthly availability with a defined credit, and that credit is the sole and exclusive remedy.
- DDoS mitigation is a posture, not a guarantee. Volumetric attacks are an SLA exclusion.
- Responsibility for what you deploy — your application, your code, your users' credentials, your data classification — stays with you. The agreement says so, and so do we.
- Telephone contact reaches your account manager. It is not a technical support channel; the support desk is.
Reporting a security issue
Found something in an environment we operate?
Clients: open it through the support desk and mark it as a security report. Everyone else: use the contact form with "Security" at the start of your message. Either way it reaches a senior operator, not a queue. Please give us a reasonable window to remediate before publishing, and we will credit you if you want to be credited.
Questions reviewers ask
What procurement and security teams ask before they sign.
-
Is Stack Harbor SOC 2 certified?
No. Stack Harbor does not hold a SOC 2 report of its own and does not claim one. Our Montréal data centres include eStruxture facilities, and eStruxture states that all its facilities are ISO/IEC 27001:2022 certified and externally audited for a SOC 2 Type 2 report. Those attestations cover the physical facility layer underneath — not the operational controls we run on top. When a client pursues SOC 2, their auditor makes every attestation determination; we operate the infrastructure controls in scope and produce the operational evidence. We are an infrastructure operator, not an auditor, and we would rather say that in the first sentence than have you find it out from a reviewer.
-
We failed a security questionnaire on our hosting — what now?
Send us the questionnaire and the reviewer's comments. The items that fail hosting questionnaires are usually the same handful: no evidence that backups actually restore, no documented patch cadence, shared credentials with no access review, no retained audit logs, no written incident process, and no clear answer to where the backups physically live. Those are infrastructure-control gaps, and they are the kind of gap we close — on a readiness sprint scoped against the actual gap list, and then by operating the controls so the evidence keeps existing between reviews rather than being reconstructed the week before one. We will tell you plainly which items are ours to close and which sit with your application, your code, or your policy. Nothing in this reply is a promise that a specific review passes; that determination belongs to the reviewer.
-
Will you complete our security questionnaire or provide an evidence pack?
Yes, for the controls we operate. That means written answers and the operational evidence behind them — patch and rollback logs, access reviews, audit-log samples, and the incident timeline and postmortem for any Sev-1 in the period. For the layer below what we operate — the data centre facility — we point you to eStruxture's published attestation for its facilities rather than restating it as our own. We answer for what we operate and we say so where we do not.
-
Where is our data stored, and does it stay in Canada?
Our infrastructure runs in two data centres in Montréal. A workload placed there stays in Canada, and we document that placement. Where a design intentionally uses a US or partner region — a shard for US customers, an edge cache — that is a decision made with you at design time and recorded, not a default. One distinction matters for Quebec organizations: Quebec's privacy law treats communication of personal information outside Quebec, including to Ontario, as something that has to be assessed. If your posture requires data to remain in Quebec, where standby and backups live is a placement decision we make with you and write down.
-
Will you sign an agreement with Quebec Law 25 and PIPEDA service-provider terms?
Residency, subcontracting, breach notification, and return-or-destruction of data are written into the agreement rather than left to assurances. That is what lets your privacy officer make the determinations Quebec's Law 25 and PIPEDA require against documented facts. We do not certify a platform compliant — the client's privacy officer does. We make sure the infrastructure supports that determination with operational evidence. The example scenarios below show how that works under a hospital network's procurement review and under a public-sector framework.
-
Who is responsible for security — you or us?
Both, with a line drawn in writing. We operate the infrastructure controls in scope for your environment: hardening, patching, backups, access controls, logging, monitoring, and incident response. Responsibility for what you deploy on top — your application, your code, your users and their credentials, and the classification of your data — stays with you. The agreement says so plainly, and so do we, because the alternative is a shared-responsibility model nobody has read until the day it matters.
-
Which of the logos on your site are partnerships, and which are just technologies you use?
One commercial partnership is named on this page: Stack Harbor is a cPanel NOC partner through WebPros, the program through which we obtain and issue cPanel & WHM licences directly. It is why cPanel escalations go to the vendor rather than to a forum thread. Every other name on our technology strip — Cloudflare, CloudLinux, LiteSpeed, and the rest — is a technology in use or one we are compatible with, and the strip labels them that way. We do not describe a technology we use as a partnership.
From the blog
Guides for the decision you are making
-
Industries & Use Cases
Canadian Data Residency: What to Require From a Hosting Provider
A Canadian CTO, privacy officer or procurement lead can use this list of written requirements to hold a hosting provider to a data residency clause.
· 10 min read
-
Business Outcomes
Failed a Security Questionnaire on Hosting? The Infrastructure Evidence Buyers Expect
SaaS founders and CTOs whose enterprise deal stalled on hosting answers can see which infrastructure evidence reviewers expect, who produces it and who can help.
· 11 min read
-
Buying Guides & Vendor Selection
How to Read a Managed Hosting SLA: Uptime, Credits and Response Times
Founders, procurement leads and IT directors can read a hosting SLA clause by clause here: what the uptime figure covers, what a credit pays and which times are committed.
· 10 min read
Send the questionnaire. We'll tell you which gaps are ours.
Forward the security review or questionnaire you are working through, with the reviewer's comments. A senior operator reads it and replies with which items we can close, which need your side, and what the evidence would look like — before any scoping call.