Any tips for designing a cloud application testing strategy that ensures reliability and scalability without causing bandwidth problems?

EngineeringCloud
3k viewscircle icon1 Upvotecircle icon2 Comments
Sort by:
CTO in Media9 months ago

I would highly recommend you shift from "testing" into "observing".
Look at Dave Farley and Mike Bryzek's youtube on "Testing In Production"

https://www.youtube.com/watch?v=Nmu4URA7pSM
https://www.youtube.com/watch?v=z-ATZTUgaAo

"Testing in Production" is not waiting until production to test, but it IS about ensuring you have end-to-end monitoring in place.

You can then shift to Adapting to load to ensure reliability and scale.

e.g. If you're worried about <2s response times for your API or App webserver, monitor average response times.  If you also have your components under monitor (CPU, Memory, etc.. for your web, database, and other backend elements) you can see what causes response times to increase with load.

Now you can forecast autoscaling strategies, and scale up where needed, when needed.

IT Manager in Constructiona year ago

Hello, testing for the evaluation of which KPI?

Content you might like

What percentage of your help desk calls are password or MFA related?

< 10%36%

10-20%29%

21-30%18%

31-40%7%

41-50%4%

51-60%2%

61-70%

71-80%

> 80%1%

View Results

What is your total Infrastructure as a Service (IaaS) cloud spend per year?

Under $1m31%

$1m-$10m41%

$11m-$50m15%

$51m-$100m3%

$100m+5%

I do not use cloud computing services.2%

Other (please share below!)

View Results

We are wondering if other companies that are hosted on the Cloud (AZURE/GCP or AWS) have implemented File Integrity Monitoring solution on Openshift? Would you have a return on experience?

Has anyone used Denodo to create an abstraction layer that contains semantic models driven by federated queries?  If so, what has your experience been related to performance? Cost?

Looking at Databricks as an option, we've been advised that Databricks is cloud agnostic, but does seem used quite a lot in cloud native services? This would suggest that any Databricks solution is not easily portable between cloud providers. Does anyone have any experience/knowledge on how easy (or not) it is to migrate Databricks between cloud providers?