Service level agreement

1.What this agreement covers

This agreement is part of our terms of service. Where your quote sets a different service level, the quote applies.

• Servers we host: managed and self-managed VPS, cloud VPS, dedicated and bare-metal servers, and GPU servers.
• Managed services: what we manage on those servers, and on managed Kubernetes, as set out in each plan's scope.
• Business email, AWS and Azure: the vendor's own service level applies to its platform. We cover our part: set-up, support, and raising problems with the vendor for you.

2.Network uptime

Our network will be available 99.9% of the time in each calendar month. Available means our network in the data centre is passing traffic to and from your server.

On managed plans we also watch the server itself: it is checked at 60-second intervals, and a failed check alerts an engineer.

3.What doesn't count as downtime

• Scheduled maintenance that we've announced at least 48 hours in advance
• Problems caused by your software, settings or content, or by changes you or your developers make
• Faults outside our network: the wider internet, your own internet connection, or another company's service
• Outages at Google, Microsoft, Rediff, AWS or Azure, which their own service levels cover
• Attacks aimed at your service, while we work to contain them
• A service suspended under our terms or acceptable use policy
• Events beyond reasonable control, such as natural disasters or a government order

4.Support response and fix targets

A first response means an engineer has read your request and replied with what they're doing. Severity 2–4 targets run in support hours, 9 am – 7 pm IST, Monday to Saturday. Fix targets are targets, not guarantees: a failed part or an upstream fault takes the time it takes, and we tell you so. You can escalate at any time — see the escalation path.

• Severity 1, down — your site, server or email is unreachable for everyone, or data is at risk: first response within 10 minutes, 24/7; updates every 30 minutes; target to restore service within 4 hours, unless a hardware fault needs the data centre.
• Severity 2, degraded — working, but slow, partly broken, or backups failing: first response within 1 hour; target 1 working day.
• Severity 3, request — a change or a task: first response within 4 working hours; done within 2 working days.
• Severity 4, question — advice, billing or a quote: first response within 1 working day.

5.Hardware replacement

On dedicated and bare-metal servers, we replace a failed disk, power supply or memory module. Most part replacements are done within 4 hours; some depend on the data centre.

6.Service credits

If network uptime falls below our commitment in a month, you get a service credit on the next invoice.

To claim a credit, write to us within 30 days of the end of the month, with the dates and times affected. Credits go on your next invoice, can't be paid out in cash, and in any month are capped at that month's fee for the affected service.

Service credits are the only compensation under this agreement for missing a target. They don't apply to the exclusions in section 3, or while an invoice is overdue.

7.Incident reports

After every Severity 1 incident, we send a written report within 2 working days: what failed, what we did, and what changes so it doesn't happen again.

8.Your part

• Keep your contact details current, so we can reach you during an incident.
• Give us the access we need to investigate, and tell us about changes that could affect the service.
• On self-managed plans, keep your software up to date and keep your own backups — see our backup policy.

9.Changes to this agreement

When we change this agreement we update the date at the top and tell customers in advance. A change never lowers a commitment during a billing period you've already paid for.