Sfn triggered twice even if only test invoke cronjob once

[rivernews]

Turns out cronjob may just only invoked once.

But the metadata S3 trigger called twice:

Interval around 40 seconds, actually quite a while. What is that metadata trigger? We should able to look at which metadata.json triggered each, closer lookout their cloudWatch:

• 23:36:13 first 09429cf6-8fc4-47b9-9cef-e3681a55434c triggered by daily-headlines/2021-08-20T22:34:47Z/metadata.json
• 23:36:57 second b16ec1d8-0868-4adc-9f63-63489ad36d69 triggered by daily-headlines/2021-08-20T23:13:25Z/metadata.json

Recall our cronjob has Saved landing page metadata to s3://**/daily-headlines/2021-08-20T22:34:47Z/metadata.json, so that's the first one no problem.

The question is why the second get triggered. We may look at its events and figure out why the trigger:

• Landing page S3 key **/daily-headlines/2021-08-20T23:13:25Z/landing.html, uuid 18acaf47-bba9-4468-9486-0adaa35ca457.
• Its LANDING_METADATA_DONE is evented at 2022-09-30T23:36:55Z. This is the cronjob outcome.
  • Cronjob log. Test invoke 7c2bce03-2c84-4126-8746-04938ddf103f done at --:36:11 -> generated metadata.json -> metadata trigger at 23:36:13. Make sense.
  • --:36:46 cronjob invoked again 9eaf1695-aeb2-4338-8871-43391530929f. Why? This is not our manual Test invoke. Only interpretation is our cronjob rate(40min) is up and happened to trigger here.
• LANDING_METADATA_REQUESTED is evented at 2022-09-30T22:44:52Z, which will always happens the same time with LANDING_PAGE_FETCHED.. so it means landing page trigger invoked, means landing page fetched. Irrelevant to our context here we can exclude it.

[rivernews]

If we're not ready, we shouldn't enable cronjob. Disable cronjob rate for now.