Blog & News

Weekly Report

API Ratings: Dec 7-14 | Pager Duty Outliers

Dec 17, 20203 min read
PC

Written by

Paul Cray

Co-Founder

Paul co-founded APIContext (formerly APImetrics), building the synthetic monitoring and conformance testing platform from its earliest days.

API Ratings: Performance

For our API ratings, we look at over 100 APIs and pull all the metrics together to give you a general feeling for the service quality for an organization’s APIs in a particular API ratings category with our patented CASC score.

Something of interest

Pager Duty has a lot of outliers – 7.10% - more than twice the number of Slack (3.38%). If we look at Total Time API ratings, we can see there is a lot of variation between locations. api ratings outliers 1 AWS North America is more than four times quicker than Azure South America and we see that the 95th percent time is nearly three times faster. This indicates that we may get a lot of the outliers from South America. api ratings outliers 2 We can see that it is the PageDuty List Incidents endpoint that is pulling down PagerDuty overall in terms of outliers. This endpoint has 18.4% outliers. Digging down we can that there is quite a big variation between clouds in terms of percentage of outliers with IBM Cloud being nearly three times more likely to produce an outlier than Azure. That’s potentially going to have a big impact on user experience. This is why you need to actively monitor your APIs from a range of clouds and locations. Just because an API works fine from one cloud location does not mean that your users who are making calls from other ones are going to be getting the same performance. api ratings outliers 3 50% of calls from South America are outliers this week, which is a serious deterioration from the previous week, but over 36% outliers is still not good. These results suggest that Pager Duty need to take a good hard look at the configuration around this endpoint in particular and especially whether there are any issues around it for IBM Cloud and South America. We don’t have any IBM Cloud locations in South America, so that can’t be the worst over cloud location. api ratings outliers 4 There are always going to variations in performance between cloud locations because the world is large and the speed of light finite, but the goal of API engineering should be minimizing them as much as possible. Actively monitoring your APIs is a vital step in ensuring that all your users get the performance they expect.

API Ratings: Overall Quality

Week Ending 14 December 2020

Category

Organization

CASC score

Corporate Infrastructure

Pivotal Tracker

9.55

COVID-19

nubentos.com

See what your APIs look like from the outside.

APIContext gives engineering, product, and customer success teams a shared view of API reliability, conformance, and customer impact — without rebuilding dashboards.

Start free