Open scbedd opened 4 years ago
Adding the header is not enough, we also need to understand where the current data is coming from a db why does it look like this.
Additionally, @pakrym suggested that we log the referrer headers to understand where our traffic is coming from. This could give us some clarity on how how github is caching the images.
Will be adding that along with cache-control.
It's not complicated to add this, and if it results in even slightly better numbers I'm completely fine with it.
Reference for Adding Cache-Control Headers
<Snipped from a mail from @pakrym >
Looking at https://github.com/Azure/azure-sdk-for-net/pull/8559 PR to update pixel tracking links today I investigated a little bit to see how exactly it works.
What if found is that Github is seems to be caching images according to https://help.github.com/en/github/authenticating-to-github/about-anonymized-image-urls and what I see in the browser so I’m not sure how pixel tracking would work.
What’s interesting is that we did collect some data from pixel-server. But the data we collected looks very suspicious.