Fundraising tech/Message queues
This page gives an overview of the message queues used to decouple fundraising subsystems. For a description of the message formats, see "Normalized donation messages". See also the Wikitech article on WMF-specific configuration.
Queues are used to decouple the payments frontend from the CiviCRM server. This is important for several reasonsː it allows us to continue accepting donations even if the backend servers are down, it keeps our private database more secure, and it enforces write-only communication from the payments cluster.
The main data flow is over the donations queue. Completed payment transactions are encoded as JSON and sent over the wire, to be consumed by the queue2civicrm Drupal module and recorded in the CiviCRM database.
Another important queue is the pending queue, which pipes messages to the pending table. Before we redirect the donor to a payment processor hosted page or iframe, we record the donor's personal information and push it to the pending queue, where it's consumed to a table indexed by gateway name and transaction ID. We store this information in a temporary fashion rather out of concern for storing data about people who aren't donors. When the donation is entered into the Civi database, we search for a corresponding pending message. We delete the message, and merge this information into the completed donation message sent to the regular queue. We also delete pending table rows when we receive notification that a payment has failed.
However, if control is never returned, then pending db messages will sit around for some time in case the data will become useful again. After about 20 minutes, they become eligible for orphan rectification, currently only applied to Ingenico credit card transactions and PayPal Express Checkout Payments. We attempt to complete settlement on these orders, and if successful, the completed message including pending-provided details is sent to the donations queue. If unsuccessful, the personal information should be purged. All pending information is purged after at most 30 days.
|donations||payments, audit processors, IPN listeners, queue job runners||DonationsQueueConsumer (queue2civicrm)||Primary queue for incoming donations, written to whenever we learn about a successful payment.|
|recurring||audit processors, queue job runners||RecurringQueueConsumer (queue2civicrm)||Information about changes in donor monthly subscriptions. Also contains information about individual recurring payments, but we should send those to donations.|
|refund||PayPal queue job runner||RefundQueueConsumer (queue2civicrm)||incoming IPN notifications that the processor has completed a refund|
|pending||payments||PendingQueueConsumer (SmashPig)||Donation methods which use an iframe (GC, Adyen) will leave a message on the pending queue before transferring UI flow to the processor. SmashPig pending queue consumer dumps this to temporary storage used by some IPN listeners and frontends. Messages will either be read from the db table after a fixed time, to complete settlement steps; or upon incoming notification of a status change; or will expire in a short amount of time.|
|payments-init||payments||PaymentsInitQueueConsumer (queue2civicrm)||One entry for each transaction that leaves our flow control.|
|payments-antifraud||payments||AntifraudQueueConsumer (queue2civicrm)||Data on transaction fraud scores|
|unsubscribe||unsubscribe page||UnsubscribeQueueConsumer (queue2civicrm)||The FundraisingEmailUnsubscribe Mediawiki extension allows donors to opt out of bulk mailings and sends these requests over a queue, to be consumed by the CRM.|
|opt-in||opt-in page||OptInQueueConsumer (queue2civicrm)||The FundraisingEmailUnsubscribe Mediawiki extension also allows donors to opt in to bulk mailings and sends these requests over a queue, to be consumed by the CRM.|
|banner-history||payments||BannerHistoryQueueConsumer (queue2civicrm)||Used to capture correlations between contribution tracking ID and banner history logging ID.|
|jobs-adyen||Adyen IPN listener||Adyen queue job runner (SmashPig)||Job types sent: DownloadReportJob, ProcessCaptureRequestJob, RecordCaptureJob. When we recieve IPN messages confirming authorizations and captures of credit card payments or availability of new audit reports we send jobs to this queue. Note that similar scaling problems to the initial paypal queue job design exist here and we may have to proliferate this queue to jobs-adyen-N.|
|jobs-amazon||Amazon IPN listener||Amazon queue job runner (SmashPig)||Job types sent: RecordPaymentJob. IPN messages from Amazon come in with only a small amount of donor information. The listener drops RecordPaymentJobs in the queue, which when run by the queue job runner, combine the payment status update with data from the pending table and drop a message in the donations queue.|
|jobs-paypal||PayPal IPN listener||PayPal queue job runner (SmashPig)||Job types sent: PayPal\Job. Originally this Job called the PayPal message verification endpoint, but we found that didn't scale and moved the verification back up to the initial IPN handler. Now this job just normalizes the IPN and drops it into one of the other queues.|
|jobs-ingenico||Ingenico audit processor||Ingenico queue job runner (SmashPig)||Job types sent: TokenizeRecurringJob. If somehow a recurring donation gets set up but not tokenized, and we see it in the audit processor, this job makes the API calls to tokenize the donation and then send it to the donations queue.|
|contribution_tracking||payments, crm-queue2civi||crm-queue2civi, analytics||Hybrid write-only log and analytic store. This is a table in the drupal db, but we're working to turn this into a queue.|
|error logging (completion data)||payments||crm-audit||Syslog, but also mined to reconstruct missing transaction data. TODO: standardize line format, use normalized data and not gateway XML.|
|banner-impressions||wmf-varnish||legacy-bannerimpressions-job||TODO: Use Kafka client without filesystem layer|
|DI-gateway-generic||push||payments-init||Donation collection frontend, which can create successful donation messages.|
|push||pending||When processing becomes asynchronous or risky, a copy of the transaction is pushed to the temporary "pending" queue, usually waiting to integrate some response from a payment provider.|
|get sequence||contribution-tracking||(future) This is a table right now, and the sequence comes from an autoincrement. It's the last link making our frontend depend on the backend db.|
|push||contribution_tracking||(future) This is a table right now, and we write directly to it.|
|crm-audit, py-audit||push||donations, recurring, refund||Nightly or weekly confirmation of payment events, in case our other methods of recording them fail.|
|crm-audit||search by id||error_logs (text files, not a queue)||This funky feature is to pull otherwise unconsumed information about failed transactions back into memory, to gather context for donations we otherwise missed.|
|Adyen IPN listener||push||jobs-adyen|
|Adyen ProcessCaptureJob (SmashPig)||push||payments-antifraud|
|Adyen RecordCaptureJob (SmashPig)||push||donations|
|BannerHistoryQueueConsumer (queue2civicrm)||pop||banner-history||Import banner history log-donation id correlations.|
|DonationsQueueConsumer (queue2civicrm)||pop||donations||The main consumer is the queue2civi job, which reads from the donations queue and stores in our CRM database.|
|AntifraudQueueConsumer (queue2civicrm)||pop||payments-antifraud||Import statistics|
|PaymentsInitQueueConsumer (queue2civicrm)||pop||payments-init||Import statistics|
|PendingQueueConsumer (SmashPig)||pop||pending||Reads messages from pending queue and dumps them in pending table in smashpig database.|
|RequeueDelayedMessages (SmashPig)||push||*||SmashPig maintenance script to move messages from the damaged table back to their original queue if they have a retry_date in the past.|
|RecurringQueueConsumer (queue2civicrm)||pop||recurring||Q2C module. Writes information about new or changed subscriptions to civicrm_contribution_recur and records payment tokens in civicrm_payment_token. Also records individual donations in a subscription, but we should leave that to DonationsQueueConsumer.|
|RefundQueueConsumer (queue2civicrm)||pop||refund||Q2C module to record refund notifications, marking the affected contribution|
|UnsubscribeQueueConsumer (queue2civicrm)||pop||unsubscribe||Decouples the mailing list unsubscribe UI from the CRM database.|
|DI-recurring-ingenico||push||payments-init||Make monthly charges. TODO: Should we push to the donations queue as well, and even the recurring events stream if we failban?|
|analytics||query by all indexes||contribution_tracking||TODO: should be decoupled from frontend event log|
See the main article: Fundraising Queue Overhaul.