Testability and application logs

30 days of Testability, day two

UneeQ

I’m not sure if this has come up on the blog yet - but I work for UneeQ, we make digital humans.

I thought a good experiment for day two would be to make one of our digital humans do ‘something’, then try and find it in the logs.

To do this, I paired up with one of our developers, Darragh McCarthy, and recorded the resulting conversation.

---

Security, privacy and that

This was a great conversation, but - we witnessed a lot of stuff that is kinda sensitive to the company, or shouldn’t be shared.

So - I’ve had to edit or censor a lot of this conversation. That’s why it’s taken so long to publish.

But, I still think it’s interesting, and you might find it valuable. So - apologies for the quality and the chopping and cutting - but enjoy anyway!

---

Learnings

I learned about using our cloudwatch logs a bit better, and how to read through a ‘conversation’ happening in our logs. It also led to some really good conversation about what is / isn’t disclosable to the public.

It probably wasn’t clear in the video, but, it raises an interesting point about when privacy clashes with testability too. For example, it would help testability if we could see all the transcripts logged there in Cloudwatch.

However, this presents a privacy problem - we can’t control what someone says to the avatar, and if that conversation is saved and logged somewhere - we could end up holding on to information we don’t want.

So - in a situation like that, privacy has to win out!

Thanks, see you shortly for day three!