We strive for as close to 100% uptime as possible.
We’ve even had entire years go by without a single downtime. Honestly, it’s not that hard. We pick the most reliable hardware vendors in the industry and take great care in maintaining our servers. All changes and updates carefully logged, no rogue patching or risky updates in the middle of the business day!
Don’t get us wrong. Downtimes can (and do) happen. The difference is being proactive about it and staying transparent with our clients. We choose the best vendors, don’t overcrowd the server, and do our maintenance carefully. Any downtimes or issues will be noted on our blog! We promise not to ghost on your tickets only to respond 2 days later with, “Everything fixed now. Try again.” We can’t sleep either when our servers are down.
The only 4 downtime scenarios you’ll ever see from us:
- Routine maintenance – typical server maintenance occurring weekly or monthly with no downtimes at all! Usually done around late Friday night or early Sunday morning. Major security flaws are patched immediately. (Expected downtime: usually unnoticeable or up to a few minutes.)
- Hardware update – only happens every 3 to 5 years, servers are taken down for about 15 minutes while being upgraded to new hardware. We’ll notify you in advanced and try to do it during low-traffic hours. In most cases, you won’t even notice.
- Hard drive failure – hard drives do fail sometimes (very rare). Not to worry, our physical machines have RAID setup and no information is ever lost. The system may be down for as much as 45 minutes while the hard drive is replaced.
- Datacenter power outage – unfortunate industry event that happens once every 5 years or so. It happens to even the biggest companies (Amazon, Facebook, etc). (Expected downtime: the very last one was about 45 minutes. The previous one was 10 years ago and about 2 hours.)
We’re not reckless with our maintenance and also not one of those companies that avoid downtime by never maintaining the server!
We use the highest industry-level TIER-4 datacenters with industry-standard 99.9% service level agreements (SLA). Sure, some promise 100% or 99.9999% SLA but it usually only means they give you more money back for downtimes (no actual difference in infrastructure).
- Again, our downtimes usually come only from rare datacenter power outages and not because of reckless maintenance or “unexplained server accidents”.
- We do offer HA-proxy for particular clients who really need it, just ask.