dionakra / gog-russian-prices Goto Github PK

The data shown in this web is obtained exploiting GOG's REST API six times a day, scheduled as a Firebase Function. That means that obtaining the data can fail due to several reasons.

1. GOG is under heavy use and its API is taking too long to deliver data. This would crash the data retrieval.
2. GOG is under heavy use, so requests take longer than normal. Firebase has a hard limit of resources usage of 600 seconds. That means 10 minutes. As I am throttling the requests so the server is not overloaded, the software is run on a Google instance with limited resources and the number of requests is pretty high (around 40), slow requests could cause the Firebase Function running for more than 10 minutes, so the function is aborted.

Because of that, I never really know if the data is refreshed or not. Well, I could know looking at the logs and executions, but no potential user could know it.

So, as we are already storing the last time a Firebase Function execution was successful under the lastRun field on Firebase RealTime Database...

• Print the last time the content of the page was updated.

The web displays in the landing page the games that are on sale right now in GOG.com. This data is retrieved from Firebase from a separate collection that holds only data for those games currently on sale.

But after retrieving the data and displaying it, the web also retries all the games, whether they are on sale or not, so the user can search and find any game, but I think most users don't really use the search option. They only look at the games on sale.

Also, this retrieval is expensive, as all the data for all the games has around 1MB in size, which is very big. So the best option here I think is to defer the loading only if he user searches for everything.

• Defer all games loading when a user start searching something.