Optimising an AWS microservice - part one #250
Conversation
| * This is stored in a DynamoDB table fronted by an API that allows the traders to PUT opinion and our system to GET it when required | ||
|
|
||
| ### The Pricing Model | ||
| * This is a java lambda - you pass it a description of what might happen in a football match along with the necessary opinion and it returns a price. |
|
|
||
| ### The Pricing Model | ||
| * This is a java lambda - you pass it a description of what might happen in a football match along with the necessary opinion and it returns a price. | ||
| * E.G. you might request “Leeds Utd to score 3+ goals” and it might tell you this is 2/1 |
There was a problem hiding this comment.
It's "e.g.", but because it's at the start of a sentence, it should be "E.g." ;-) I'd just opt for spelling it out though:
| * E.G. you might request “Leeds Utd to score 3+ goals” and it might tell you this is 2/1 | |
| * For example, you might request “Leeds Utd to score 3+ goals” and it might tell you this is 2/1 |
|
|
||
| ### Pricing Service | ||
| * This is made up of an API Gateway and Java/Kotlin lambda | ||
| * The API Gateway has a ‘Lambda Authoriser’ that ensures our system can only be accessed from whitelisted ip ranges |
|
|
||
| ## Our Test Scenario | ||
|
|
||
| We wrote a performance test in Node JS that populates our opinion store with fake fixtures and then sends requests to our pricing service, recording the latency of each request. |
There was a problem hiding this comment.
We tend to use "NodeJS" as the term everywhere else
There was a problem hiding this comment.
For consistency, we should match the canonical form used by the maintainer; our most common ones are listed on the jobs page, so in this instance we should use "Node.js"
|
|
||
| ## Uh oh… things aren’t as they seem :-( | ||
| Now under our revised target, we were about to wrap up the project when someone noticed something odd. Our DynamoDB table only had opinion for 1 fixture in despite the test saying it had created 100 :scream: Uh oh! | ||
| It turns out we had a typo in our node code in the performance test which had slipped through review. |
|
I'd generally advise to remove the smart formatting (e.g. curly quotes, en dashes, ellipses) and have Markdown apply that automatically |
|
This is my first attempt at a blog post so hopefully I've got it right...
Have loaded using docker and had a quick review - all seems to render correctly.
There is a part 2 written, will submit another PR for that separately.