Understanding uptime guarantees: what 99.9% SLA actually means for Indian websites

99.9% uptime sounds almost perfect. It allows 8.7 hours of downtime per year. Here is what that really means for your business and how to evaluate claims.

The maths of uptime

Most shared hosting providers offer 99.9%. Most managed VPS providers offer 99.95% or 99.99%.

"99.9% uptime guaranteed" is the industry minimum, not a claim of excellence. And most providers hit it — but how they define "uptime" and what they do when they miss it matters significantly.

What counts as downtime

This is where SLAs become tricky. Read your hosting contract for the definition.

What most providers count:

• Server completely unreachable (ping fails, no HTTP response)

What many providers exclude:

• Scheduled maintenance windows (some providers carve out 4 hours/month)
• Issues caused by your application code or plugin
• DDoS attacks
• "Force majeure" events (power outages, natural disasters)
• DNS propagation issues
• Third-party service failures (payment gateways, email providers)

A scheduled maintenance window of 4 hours/month already accounts for most of a 99.9% SLA's allowed downtime. Some providers are effectively guaranteeing nothing.

What to look for in the contract:

• Is scheduled maintenance excluded from the SLA calculation?
• What constitutes a "confirmed outage" vs "intermittent connectivity"?
• Is the measurement period monthly or annual?
• Who decides if an outage actually occurred?

What SLA credit actually means

Most hosts offer SLA credit — a percentage of your monthly fee back if they miss their uptime target.

Typical credit structure:

• 99.9–99.0%: 10% credit
• Below 99%: 25% credit
• Below 95%: 50% credit

On a ₹200/month plan, a 25% credit is ₹50. This is roughly what an hour of downtime costs the host to compensate you — regardless of what an hour of downtime actually cost your business.

The mismatch: An e-commerce site going down during a sale might lose ₹50,000 in an hour. The SLA credit is ₹50. SLA credits are not compensation — they are a gesture. Do not choose a host based on SLA credit generosity; choose based on actual historical uptime.

How to check real uptime before you commit

Look for public status pages

Good hosts maintain public status pages (like status.villagehosting.in) that show real-time and historical availability. Check the history tab — is it genuinely green for months, or does it show regular incidents?

Ask for uptime percentages for specific plans

Ask the sales team: "What was the actual measured uptime for your [specific plan] in the last 12 months?" A confident answer with numbers is a good sign. Vague assurances are not.

Search for hosting reviews with outage mentions

Search "[host name] outage 2025" or "[host name] down" on Twitter, hosting forums, and Reddit. The pattern of outage reports (how frequent, how long, how the host communicated) tells you more than the SLA.

Monitor independently after signing up

Set up UptimeRobot (free) or Freshping to monitor your site from outside your host's network. Check from at least two locations — one in India, one international.

If you measure lower uptime than what the SLA promises, document the incidents with timestamps and request credit.

What causes downtime for Indian hosting

Shared hosting: noisy neighbours

On shared hosting, one customer running a poorly-coded script or getting a traffic spike can affect other accounts on the same server. Good hosts isolate resources with CloudLinux; bad hosts let high-CPU scripts degrade the entire server.

Ask: Does your shared hosting use CloudLinux or cgroups for resource isolation?

VPS: overselling

VPS providers oversell — they sell more virtual resources than the physical server has, betting that not all customers use their allocation simultaneously. When many customers are active at the same time (often during Indian business hours: 10am–7pm IST), oversold servers run out of resources.

Ask: What is your physical-to-virtual overcommit ratio for CPU and RAM?

Infrastructure: single points of failure

A hosting provider with one data centre and one ISP upstream link is one cable cut or power event away from a significant outage. Redundant power (multiple UPS units, diesel generator), redundant network (multiple ISP connections), and ideally a secondary location for failover separates serious infrastructure from commodity hosting.

What "99.99% uptime" actually requires

To reliably achieve 99.99% uptime (4.4 minutes/month), a host needs:

• Redundant power (N+1 UPS, generator)
• Redundant network (at least 2 upstream ISP connections)
• RAID storage (multiple disks so one failure does not take down the server)
• Hot standby or load-balanced server configurations
• 24/7 proactive monitoring (not reactive — the team knows about a problem before customers report it)
• Tested failover procedures

This is achievable on dedicated server and managed cloud infrastructure. It is not realistic for most shared hosting setups.

What to do when your site goes down

1. Verify it is not just you: visit downforeveryoneorjustme.com
2. Check your host's status page
3. Check your monitoring tool (UptimeRobot) for when it started
4. Contact support — phone or WhatsApp if available, they should know already
5. Document the outage with timestamps for SLA credit claims
6. Check if the issue is on your server or the host's network by pinging the server IP directly

Our uptime commitment

VillageHosting maintains a public status page at status.villagehosting.in. Our infrastructure includes redundant power, dual upstream connectivity, and NVMe RAID storage. Managed plans include proactive monitoring — our team is notified before customers.

We publish real incident history on our status page, including the cause and resolution for every incident. If we miss our SLA, we issue credit automatically without requiring you to submit a claim.