Use delayed job push for RecordLintJob

Why:

• RecordLintJob is currently enqueued eagerly via JobQueueGroup::push().
• Since this also occurs on pageviews with Parsoid enabled, it's taking up ~1.1% of total request wall time according to Excimer data from Jan 2nd.[1]
• We don't want the request to fail if the job cannot be enqueued, so we can safely use lazyPush() instead to allow this job to be batched and enqueued as part of a post-send deferred update.

What:

• Use lazyPush() instead of push() to enqueue RecordLintJobs.
• Remove the now-redundant error handling wrapping the enqueue.

[1] https://performance.wikimedia.org/arclamp/svgs/daily/2025-01-02.excimer-k8s-wall.all.svgz?s=onParserLogLinter

Bug: T382975
Change-Id: I5df7e63e497fe509096200be68b38e4888d9dc02