What happens when an ESP data center goes down?

Your ESP's status page goes orange. Something's down. Do you panic, or trust the system? The answer mostly depends on how well your ESP built their infrastructure in the first place.

Most major ESPs run across multiple data centers in different regions. If one goes down, here's what's supposed to happen automatically. Load balancers detect the failure (usually within seconds) and stop routing traffic to that location. New messages get handed off to healthy data centers instead. Your sending continues, maybe with a slight delay, but it keeps going.

The emails already sitting in queue at the failed data center are the trickier part. Good ESPs store queued messages on persistent storage, so nothing gets lost when a node goes dark. Once that data center comes back online, those messages drain normally. Some ESPs also keep redundant queue copies in other regions so even that gap closes fast.

What you might actually notice depends on how well-prepared your ESP is.

• Well-prepared ESP: slightly higher latency for a window, maybe nothing at all. Dashboard might show a blip.
• Poorly-prepared ESP: sending delays, queues backing up visibly, possible message loss, and a dashboard that also stops working (which is the worst combination).

Once engineers bring the failed data center back up and verify it's healthy, traffic gradually rebalances. Any stuck queues drain. Things return to normal, and most senders never knew anything happened.

The honest takeaway is that a single data center outage shouldn't be a crisis for you. It should be a crisis for your ESP's ops team. If your ESP doesn't have multi-region redundancy, though, that's worth knowing before you're mid-send on a time-sensitive campaign. It's worth asking your ESP directly about their disaster recovery setup before you actually need it.

Not sure what questions to even ask? Drop us a line at our SOS hotline and we'll help you figure out what your ESP should (and shouldn't) be doing.