[APM] Normalize time range for apm-per-service telemetry

[basepi]

Summary

After discussions on #144061 I think we should probably normalize apm-per-service to also use the range1d filter instead of now-1h.

Edit: At @graphaelli's recommendation I'm going to have it do 1d, check the timeout, and if it timed out do the 1h version. I'll re-request review at that point.

For maintainers

• This was checked for breaking API changes and was labeled appropriately

[elasticmachine]

Pinging @elastic/apm-ui (Team:APM)

[graphaelli]

could this try first for 1d and back off to 1h in case of timeout?

[kibana-ci]

💚 Build Succeeded

• Buildkite Build
• Commit: bc32a16

Metrics [docs]

✅ unchanged

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream

[basepi]

could this try first for 1d and back off to 1h in case of timeout?

Good idea! Since we're using proper timeouts, running the queries twice shouldn't be much of a problem. I'll move this to draft and make a note in the description, I don't have time to do this properly at the moment.