There are plenty of API metrics that can be generated – but what are they really telling you?
APImetrics has spent years focused on the key questions about API performance and bringing AI based reporting to API monitoring in a single monitoring and reporting product.. This is so you can identify what you need to know versus what might mislead you into thinking all is well, when, in fact, it isn’t.
To achieve this, APImetrics created a database of over a billion API calls, and 10 billion data points in over 30 TB of data. So not only do provide you with state of the art monitoring for latency and uptime – but you can also identify data you may be missing, like the quality of the overall service and where the pain points are.
This article outlines the key API metrics we store and the types of reporting you can use to understand them better.
AI Driven Insights
At the heart of APImetrics is the AI Based API Insights Report. The analysis works from the calls you make and the training our system has gained from the billions of data points we’ve collected since we started monitoring APIs globally in 2016.
The data that goes into Insights is based on:
- Pass Rates – as measured by the criteria you set in your API tests
- Latency – the time it takes the API call to execute
- Outlier rates – calls which the AI determines fall outside of your normal test criteria
- Consistent response times – your API may be blisteringly fast, or snail-like slow; do you have periods of time when it’s much slower than you expect?
- Overall speed – we don’t look too hard at the speed, but we do keep an eye on a famous IT metric of about 450ms (the speed IBM determined that a computer system had to respond before people noticed the lag)
These API metrics are pulled together in the CASC score, which tells you out of a perfect 10 if there are any problems with a specific API. The clever thing with CASC is it allows us to compare different APIs, so you can see how they rate against each other or a competitor.
Regional metrics are generated so you can see Cloud, Location, and Regional API performance metrics at a glance to understand where there might be problems. This provides insight to internal engineering AND external customer facing teams.
What API Insights Tell You
With the AI Based API Insights report, you can learn which data centers and clouds are providing the worst service – a critical piece of data for informing third parties about their issues.
If you know, for example, that AWS Frankfurt is 2 seconds per transaction slower than Azure London, then you’re equipped to understand why people might be complaining about a specific location.
Likewise if you’re using containers, you might want to know where in the world and on which of the major clouds you get the best performance.
APImetrics can do that and more. We also provide API metrics benchmarking so we can easily show how your specific API ranks against groups of APIs, industries, and more.
Integrated Statistical Analysis Tools
With APImetrics you don’t have to export your data into yet another product. With our fully integrated statistical analysis system, you get all the data you need in our statistics views and reports – without having to sign up for, or pay for, another product.
Statistics analysis gives you data on the actual performance metrics for your APIs for as long as you’ve been a customer. This includes:
- A geographic mapping feature to show calls by call segment (DNS, TCP, Processing Times) and Cloud on an interactive world map view including percentile breakdowns
- Pass/fail rates
- Latency graphs for all call segments and also the ability to map different APIs together for direct comparison purposes – these can be saved for future reference
- Heatmaps to show when calls are slow either over an hourly/daily/weekly/monthly basis so you can visualize performance trends for the speed of your responses
Reporting and Status Pages
In addition to statistical analysis tools we provide a range of reporting and status page options – including the ability to download reports as a CSV file.
Reports show API pass rates either as a total or by call location, and break it down further by the data center that they were made from. Data is also presented with 95th and 99th percentile analysis so you can see which APIs have long extreme times that can impact you.
Every report can also be automatically turned into a real-time status page for direct embedding into a blog, operations center, or other similar reporting service for internal and external users.
Compliance and Governance
Imagine being able to generate the reports your regulators need automatically – from real external performance data – just like users of your platform experience.
Imagine being able to bring that data to dispute resolution discussions with actual evidence that it isn’t your platform, or show where the weak links are in somebody else’s stack.
Well, imagine no more. APImetrics provides automatically generated regulator reports for the key regulated industries or we can customize the output of the product to generate exactly what you need when you need it.
Coming soon to APImetrics – integrated SLA and SLO management and automatic Service Level tracking.
Ready To Start Monitoring?
Want to learn more? Check out our technical knowledge base, or our sector by sector data, or even our starters guide to the API economy. So sign up immediately, without a credit card and be running your first API call in minutes.