RUM-1507 Improved diffing

[maciejburda]

What and why?

It leverages new resourceId field instead of base64 for calculating diff of image wireframes, speeds up the calculation and lowers the CPU pressure.

How?

We added new field to SR schema (that will be later on used for different purposes) and started using this field instead of base64 to calculate object's hash.

Now, on Shopist it goes as low as 12% of CPU for the most complex view on idle. Used to be 60% before the change.

Review checklist

• Feature or bugfix MUST have appropriate tests (unit, integration)
• Make sure each commit and the PR mention the Issue number or JIRA reference
• Add CHANGELOG entry for user facing changes

Custom CI job configuration (optional)

• Run unit tests for Core, RUM, Trace, Logs, CR and WVT
• Run unit tests for Session Replay
• Run integration tests
• Run smoke tests
• Run tests for tools/

[datadog-datadog-prod-us1]

Datadog Report

Branch report: maciey/RUM-1507-resource-identifier
Commit report: b0fc4b7

✅ dd-sdk-ios: 0 Failed, 0 New Flaky, 241 Passed, 0 Skipped, 4m 48.8s Wall Time

[xgouchet]

Note

Random question, do we have a benchmark on the improvement this cache is adding vs the time needed to md5 a full image data?

[maciejburda]

Hash is lazily computed once for each resource, so it shouldn't matter that much.

Optimisation of this PR happens in diffing algorithm which takes Hashable protocol into account. In it's implementation we are now ignoring base64 field.

What you're asking here is going to happen in a Resource Endpoint PR where we stop calculating base64 and upload resources directly. In this next PR, I'll measure the difference.