Shipping API Monitor – What Does 1 Year Tell Us?

One of our projects here at LexiConn – www.ShippingAPIMonitor.com – just had a birthday.  The website recently passed 1 year of monitoring the FedEx, UPS, and USPS shipping APIs, and we thought – why don’t we take a look at some of the stats and see what we’ve tracked over the last 12 months?

What Does It Do?

ShippingAPIMonitor.com provides real-time tracking of the shipping API systems for rates being returned from the big three – FedEx, UPS, and USPS. It monitors for their system being up and returning valid rates every minute from 2 geographically diverse locations. Any outage is displayed on the site, and after 3 consecutive rate failures, we mark the shipping API as down.

For this post, we’ve summarized the outages from July 1, 2013 through June 30th, 2014.

How Long & How Many?

In the 1 year of tracking we monitored a total of 1,559 minutes of downtime across all 3 APIs, which is just under a total of 26 hours.  26 hours may seem like a lot, but since that is over 1 year it’s just 0.3% of overall downtime.

Luckily most of that was at night, but you don’t want any downtime when it means a customer could fail to place an order and you therefore lose a sale.  Let’s take a closer look at our big 3:

FedEx had 173 outages comprising 1,308 minutes (21.8 hrs) of downtime.  61 of those outages totaling 1,188 minutes (19.8 hrs) were over 3 minutes long, resulting in a “Down” rating on ShippingAPIMonitor.com.

They had 6 outages last longer than an hour, with the longest being 4-1/2 hours.  That outage of was one of their planned maintenance windows, lasting from 10pm Eastern time on 8/3/2013 until 2:30am the following morning.

USPS had 130 outages comprising 250 minutes (4.2 hrs) of downtime.  18 of those outages reached our 3 minute threshold, which accounted for 123 minutes (2 hrs) of their downtime.  Their longest outage was 32 minutes, during the afternoon of 3/26/2014.

UPS had 1 outage of 1 minute in length.  That was it!  It was not even long enough to trigger a “Down” rating on our site, just a “Warning” notice for 1 minute until the next test completed as passed.

What Have We Learned?

We started this project to help our customers when they found shipping errors in their shopping carts, but we never knew how often the shipping APIs really went offline until we started tracking.  During the first month of going live (July 2013) we tracked 18 minutes of outages and then August followed with 2 scheduled maintenance outages by FedEx and a variety of smaller outages which allowed us to tweak our code when needed and confirm everything was monitoring correctly.

We quickly learned that FedEx has a lot of little outages, mostly going unnoticed by merchants since they are short and if a customer did receive a shipping error but tried again they would have been able to submit their order without ever letting the merchant know an error occurred.

USPS also has a good number of outages, but with so many being 2 minutes or less they mostly go unnoticed.

It took over 5 months before we finally tracked an outage with UPS.  What does UPS do that is different than FedEx and UPS?  We can only assume they are load-balancing their API servers with a method where at least 1 location is always (almost!) up and serving rates, as having just 1 minute of downtime in a whole year of tracking is remarkable.

Want to know when shipping APIs go down for your store?  Visit ShippingAPIMonitor.com to view stats or sign up for the newsletter to receive emails anytime an API goes down.  You can even follow us on Twitter for notifications as well.

Looking for a web host that understands ecommerce and business hosting?
Check us out today!

5 Comments

  1. Gari Anne says:

    Interesting info, maybe I should look into UPS as an option also. Thanks for this great article.

  2. Robyn says:

    Fascinating info, I find the ShippingAPIMonitor site incredibly useful in understanding down times. Also amazing that UPS is seemingly so solid and predictable.

  3. Matt says:

    I find your tool frankly wrong. We’ve been experiencing outages on UPS for the last 24 hours but according to you it’s been up. Yet we’ve had 250 “Transient 250050 License system not available” out of maybe 1000 calls.

    Further, we get these every month or so though usually only for an hour. The rest of the time it is up.

    My point is your metrics obviously don’t catch this which makes be wonder what your testing protocol is. Up versus complete transaction are two different things.

    By the way UPS support are next to useless on this. They seem oblivious their system is failing and respond with “check your code”. Cheers…

    • We are not only logging in, but requesting a rate from one zip code to another, and verifying a legitimate rate is returned. We have 1000’s of ecommerce clients that rely on UPS rates, and we have not had any reports of widespread outages.

      It sounds like something unique to your account or server/location, as we have not seen any issues…

Leave a Reply to Matt