CUNY Academic Commons

When a forum post is created by email, it is sent to our email parser provider, SparkPost. SparkPost then parses the email and sends the data back to the Commons website. When the Commons is down, SparkPost does the following: https://support.sparkpost.com/docs/tech-resources/webhook-data-streams#webhook-batch-retry-logic

• The webhook service will reattempt to post batches for which it receives a non 2xx response code.
• A timeout for any particular batch posting attempt will occur after 10 seconds if the endpoint does not successfully accept a given batch.
• The retry logic for the posting of batches gradually increases in a logarithmic fashion and will cease retrying altogether after 8 hours. This means you can potentially lose event data if your webhook consumer is completely down for 8 hours or longer. 12 total attempts will be made to POST the webhook batch to the webhook endpoint (the initial attempt + 11 retries).

It's possible that SparkPost attempted to send Luke's forum topic back for processing because the first time might have taken more than 10 seconds to post. This likely explains the second topic being posted, but the three-day delay is a little odd. I'm going to try and look into SparkPost's event logs to see if this can be explained.

It looks like SparkPost's Relay Webhooks API, which we use so SparkPost processes the emails and sends the parsed response back to the Commons, lacks some endpoints to search and query for event status compared to their regular Webhooks API. Meaning I cannot query for this information with their documented API. I've reached out to SparkPost support to see if they can provide any additional insight regarding their Relay Webhooks API. Will see what they say.