[chore] Updates to scrape duration

[MovieStoreGuy]

Description:

• This is to prepare for the future release of core that includes scraping from collector start.
• Fixes accidental change to apache receiver's default collection interval.

Link to tracking Issue:

open-telemetry/opentelemetry-collector#7635

Testing:
NA

Documentation:
NA

[andrzej-stencel]

Perhaps useful to add that the change was mistakenly added in v0.78.0?

[MovieStoreGuy]

Don't merge til open-telemetry/opentelemetry-collector#7635 is merged.

[andrzej-stencel]

I think this PR also needs to update the docs of all the affected receivers to add the new initial_delay configuration option.

[andrzej-stencel]

This is great. Now I wonder: is this a breaking change? 🤔 It changes the time of the first scrape for all the affected components - from the value of collection_interval (which is 10s or 60s for most components) to 1s. Is this change OK? Or should we rather make the default initial_delay be the same as collection_interval so that there's no change in (the default) behavior? 🤔 Perhaps it's fine, I cannot think of an issue that might arise from this change.

[MovieStoreGuy]

This is great. Now I wonder: is this a breaking change? 🤔 It changes the time of the first scrape for all the affected components - from the value of collection_interval (which is 10s or 60s for most components) to 1s. Is this change OK? Or should we rather make the default initial_delay be the same as collection_interval so that there's no change in (the default) behavior? 🤔 Perhaps it's fine, I cannot think of an issue that might arise from this change.

The interval changes would only really impact users who have larger intervals since their current expectations is to wait that to capture data. The gut feeling is that collection intervals under 1m don't notice since something like 10s over a duration of an hour is hidden is most observability tools rollups or aggregations.