Service Level
Agreement

For Cloud EA, this SLA covers the path we operate to a published-plan virtual machine: the host capacity, the attached storage we provide, and the public network we assign in Nairobi, once the instance shows as running after a cleared payment.

For retained services, this SLA covers the response of the Nairobi desk to incidents you raise on estates we have agreed to watch — networks, security stacks, voice, Microsoft tenants, or a managed VPS — during the hours on the retainer.

It does not cover a product we only resell (a Microsoft tenant, a Sophos licence, a 3CX or Yeastar licence) except to the extent our own implementation or retainer hours are on the order.

The monthly availability target for a published Cloud EA instance is 99.9%, measured over the calendar month in East Africa Time, after subtracting the exclusions below.

Availability is the time the instance is reachable on the network path we operate and is not in a failed host state we caused. It is not a warranty that your application, guest operating system, or firewall rules inside the guest will answer.

99.9% allows for up to 43 minutes of included unavailability in a 30-day month before a credit may apply.

We may take planned maintenance on the platform. Where we can, we will give at least 48 hours’ notice to the email on the portal account and we will prefer windows outside desk hours.

Emergency maintenance to protect the platform or other customers may proceed with less notice. Planned and emergency maintenance that we announce, and that is reasonably necessary, does not count against the 99.9% target.

The Nairobi desk is staffed Monday–Saturday, 8.00 AM–5.00 PM East Africa Time. Emergencies on retained estates and P1 Cloud EA incidents are covered around the clock.

Response is the time from a ticket, a monitored alert we accept, or a call we log, until an engineer is working the incident — not the time to full restoration.

• check_circle P1 — Instance unreachable or a retained production service is fully down: response within 1 hour, 24 hours a day.
• check_circle P2 — Severe degradation, a site down, or a security control failing open: response within 4 hours in desk hours; if raised after hours, by the next morning unless you mark it as an emergency.
• check_circle P3 — Standard change, how-to, or a non-urgent fault: next business day in desk hours.
• check_circle P4 — Planned work and quotations: as scheduled with the account owner.

Published Cloud EA Linux plans include daily backups stored separately from the live disk. Unless the quotation sets a tighter number, the recovery-point objective is up to 24 hours and the recovery-time objective is best effort during desk hours, or during emergency cover for a P1.

A daily backup is not an application-consistent copy of every database you run. If you need a shorter RPO, a tested restore drill, or a second site, that is a line on the quotation.

You remain responsible for copies of data that sit only inside the guest until a backup window has run.

If a published instance misses the 99.9% monthly target after exclusions, you may claim a service credit against the next invoice for that instance, as follows.

• check_circle Monthly availability below 99.9% and at or above 99.0%: 10% of that month’s recurring instance fee.
• check_circle Below 99.0% and at or above 95.0%: 20% of that month’s recurring instance fee.
• check_circle Below 95.0%: 30% of that month’s recurring instance fee.

Credits are the sole remedy for unavailability under this SLA. They are not cash, not transferable between organisations, and not an extension of term unless we apply them that way. The credit in a month will not exceed 30% of that instance’s recurring fee for that month. Add-ons (extra addresses, a managed retainer) are credited only if that add-on itself failed.

Open a ticket from the portal or write to info@patolink.com within 14 days after the month ends. Name the instance, the dates and times in East Africa Time, and what you observed. We will compare that with platform records and reply with the credit or with why it does not apply.

A claim made late, without enough detail, or for a period already credited, will be declined.

A managed-services or managed-security retainer is our time, not a second Cloud EA instance. Response follows the priorities above for the estate named on the quotation. Restoration time depends on the vendor, the access you give us, and whether a spare is on the order.

Retainer credits, if any, are written on that quotation. This public SLA does not invent a percentage credit for professional hours.

The following do not count as unavailability and do not earn a credit.

• check_circle Faults in the guest operating system, your applications, your firewall or routing inside the instance, or credentials you control.
• check_circle Suspension for unpaid fees or an Acceptable Use Policy breach.
• check_circle Denial-of-service, internet, or power events outside the path we operate, or a force-majeure event.
• check_circle Planned or emergency maintenance as described above.
• check_circle A Custom instance still being built, or a published instance not yet provisioned because payment has not cleared.
• check_circle Beta, trial, or complimentary capacity.
• check_circle Your request that we power off, rebuild, or resize the instance.

Incidents and claims: info@patolink.com.

Phone: +254 709 343 000 / +254 722 616 999.

Office: Unit 1105, Applewood Adams, Ngong Road, Nairobi, Kenya.

Hours: Monday–Saturday, 8.00 AM–5.00 PM East Africa Time. P1 cover is around the clock.