tweak default cache lengths by category

[chris48s]

Quick refresher on how we set max-age:

• Individual badge classes can set a _cacheLength specific to that service
• If the badge class itself doesn't define a _cacheLength we fall back and use the default for that category. This is the code I am changing in this PR
• If there is no default set for the category, we fall back and use defaultCacheLengthSeconds (which we set to 120 seconds in production)

The basic principles of setting defaults by category are:

• There are some data points that need change more frequently than others. Whether your build is passing or failing might change quite frequently. You rarely change your project's license.
• There are some badges that show quite a precise value (build status, version number), whereas others show a more indicative metric (number of downloads/stars/likes). These are usually rounded with metric(). It is less crucial for an indicative metric to be fresh.

In this PR I am suggesting we tweak the category defaults a bit and add some categories that were previously using the global default of 2 mins. The broad result of all of this being: Cache more things for longer. I am focussing on categories that make up a large proportion of our traffic.

We can always continue to tweak these thresholds.

Also:

• This doesn't change any situations where we've set a _cacheLength on individual badge classes - those continue to override the defaults by category
• If there are specific badges where we want to set a shorter length than the category default then we can

[github-actions]

	Messages
📖	✨ Thanks for your contribution to Shields, @chris48s!

Generated by 🚫 dangerJS against 50b194f

[chris48s]

My logic here is that the cadence with which platform support and size change is probably about once per release, so lets treat them the same as version numbers.

[jNullj]

I think that platform-support tend to change less then version and size, as most versions would not change supported platform or change major version, but its an improvement from 120s so we might just start with that.

[chris48s]

All of these categories primarily show what I've described as "indicative metrics"

[jNullj]

30min seems fair.

[jNullj]

LGTM, hope this reduces resource use with minimal effect on client usage perception