Uptime Percentages: The Actual Numbers
Uptime percentage sounds like a small difference between 99% and 99.9%. The downtime implications are not small.
| Uptime SLA | Annual Downtime | Monthly Downtime | Weekly Downtime |
|---|---|---|---|
| 99.99% ("four nines") | 52 minutes | 4.4 minutes | 1 minute |
| 99.9% ("three nines") — M A Global Network SLA | 8.7 hours | 44 minutes | 10 minutes |
| 99.5% | 43.8 hours | 3.6 hours | 50 minutes |
| 99.0% | 87.6 hours | 7.3 hours | 1.7 hours |
| 98.0% | 175 hours | 14.6 hours | 3.4 hours |
The Business Cost of Downtime
For a 10-person accounting team using cloud-hosted Tally at ₹50,000/month in combined salaries, 1 hour of complete downtime costs approximately ₹300 in direct salary (assuming 8-hour workday). But productivity loss extends beyond the downtime itself — work that could not be completed must be done later, often under pressure. The real cost of a 7-hour downtime event is closer to ₹5,000–₹15,000 in lost productivity and recovery overhead.
The difference between 99.9% SLA (8.7 hours annual downtime) and 99.0% SLA (87.6 hours) is 79 additional hours — roughly 10 full working days of potential downtime per year. At any reasonable productivity cost per hour, this difference far exceeds any price savings from choosing a lower-reliability provider.
SLA vs Actual Uptime: What to Look For
An uptime SLA is only meaningful if it includes a compensation mechanism. Three questions to ask:
- "What is the uptime SLA, and what compensation applies if you fail to meet it?" — SLA without compensation is a marketing claim, not a commitment
- "Is planned maintenance counted against the SLA?" — Some providers exclude maintenance windows from uptime calculations
- "How is uptime measured — from external monitoring or from internal logs?" — External monitoring provides independent verification
Why 99.9% Requires Proper Infrastructure
Achieving 99.9% uptime consistently requires redundant infrastructure at every layer: redundant power, redundant cooling, redundant network connections, and redundant hardware. This is only achievable in a professionally managed data centre — not in an office environment with a single UPS and one internet connection.
An office server's realistic uptime: 95–98% in a typical Indian office environment, accounting for power fluctuations, occasional UPS exhaustion, internet connectivity failures, and hardware issues. The 1–4% downtime represents 87–350 hours of potential unavailability per year.
Uptime During GST Filing Season
The timing of downtime matters as much as its total duration. 8.7 hours of scheduled maintenance distributed across Sunday nights is relatively harmless. 1 hour of downtime on the 19th of the month at 7 PM — when your team is finalising GST returns — has disproportionate impact.
M A Global Network schedules maintenance windows during off-peak hours and provides advance notice. Emergency maintenance for critical security issues may occur without advance notice but is rare. The overall approach prioritises uptime during Indian business hours and filing season peaks.
Frequently Asked Questions
Contact our support team. M A Global Network reviews downtime events and provides service credits for SLA failures caused by infrastructure issues within our control. Network issues beyond our control (internet backbone events affecting multiple providers simultaneously) and planned maintenance windows notified in advance are typically excluded. For SLA claims, document the downtime periods with timestamps and raise a support ticket within the billing period.
99.9% Uptime SLA — In Writing
Backed by redundant data centre infrastructure. ₹700/user/month + 18% GST. 7-day risk-free guarantee.