In order to retry messages that failed to send to the marketplaces, we need a new internal api to trigger sending the billable_usage kafka messages again. A cron job is also required to trigger the process.

Swatch-tally-service will get a new internal API for this initiating this:

• Billable Usage stuff is bundled as part of swatch-tally, but is kind of treated as its own thing (eventually to become its own service), so it has its own API schema that should be updated instead of the internal-tally one (internal-billing-api-spec.yaml) -> This was done as part of SWATCH-2103
• Needs PSK for auth -> This was done as part of SWATCH-2103
• Include an optional timestamp query param (as_of) to enable testing, without timestamp, assume now()
• Query billable_usage_remittance for retry_after > as_of (and non_null)
• Produce kafka message onto platform.billable_usage again
• We don’t need any additional information here 

• Clear retry_after column in the table

AC: 

• business logic for new internal API filled out that resends billable_usage kafka messages where billable_usage.retry_after < after_on.
• cron job to call the api every 1 hr

Cronjob

• This happens to identify the billable_usage_remittance records that are ready to be retried.  Retry_after < now()
• Swatch-tally ClowdApp will get a cron job definition
  •  Every 1hr
• Cronjob is calling POST endpoint, /internal/rpc/remittance/processRetries (see linked card for API definition)
• GET /internal/remittance/accountRemittances will show the retry_after field . (For QE, internal_api_account_remittance() in mqe-tools hits that end point)
•