Page MenuHomePhabricator

Donor Portal - Investigate Why Some Donors Didn't Receive Link/Explore Monitoring Options
Closed, ResolvedPublic

Description

Just flagging incase we want to look at the login for donor accounts. Typically the login links come immediately. A donor reported trouble receiving the link to login I think they might be trying the wrong email but I went just to test anyways.. A few minutes ago around 4:48 PM EST I requested my login link for my CID 68209875 and took the attached screenshot to confirm what I'd submitted. Waited 2-3 minutes or so and didn't receive the email. Requested it a second time to see what happened and got the login email immediately. Assuming that the email I received might be from the second request as it came immediately like normal, but I have no way to confirm that. Not sure if we can tell on the backend that two requests were submitted?

From today, Zendesk ticket # https://wikimedia.zendesk.com/agent/tickets/1906513 / CID 2524753 and donor notes "I have not been able to get a link to log into my donor account. I am sure of the email address: [redacted]. You have been sending correspondences to this email address. It is not in my spam account and I show no new emails from you with a search." Their donations are under the email they provided.

Also from today, Zendesk ticket # https://wikimedia.zendesk.com/agent/tickets/1906407 / CID 10300661 and the donor notes "I am unable to log into my donor account. I don't receive an access email. I've tried both my email addresses." Their donations are under the email they're writing in on and they also confirm that's an email address they tried.
From Sunday, Zendesk ticket # https://wikimedia.zendesk.com/agent/tickets/1905979 / CID 9488259 donor wrote "Hi, I have requested a login link to access my donor account THREE times but have not received it. How can I update my account?" Not sure which email they tried to use to login as they don't share that, but they do have 2 active donations under the email the wrote in under so I would think they possibly tried that one. We already replied to them today to inquire how we can assist them internally instead.

Details

Event Timeline

We found some errors in the job logs that look like a timeout in the connection from CiviCRM to the mail server. The job was configured to run for 10 minutes at a time, and the timeouts all happened at least 5 minutes after the job started. We changed the job to only run for 3 minutes at a time so it will get fresh mail server connections more often.

Change #1319523 had a related patch set uploaded (by Ejegg; author: Ejegg):

[wikimedia/fundraising/crm@master] Use custom mailer for new link emails

https://gerrit.wikimedia.org/r/1319523

Change #1319523 merged by jenkins-bot:

[wikimedia/fundraising/crm@master] Use custom mailer for new link emails

https://gerrit.wikimedia.org/r/1319523

@Ejegg: Is it possible that this was related to the time limit change here T432287: Email servers picking up same emails?

@Lars I don't think so - the job had been using the core CRM_Utils_Mail::send function so it wasn't actually connecting directly to either MX. Looks like the connection that was timing out would have been to the local postfix (?) daemon. The change in restart frequency was probably enough to avoid timeouts, but then I also deployed a change to use

MailFactory::singleton()->getMailer()->send

which uses a new connection for each mail and also respects the environment vars pointing to one MX or the other.

XenoRyet set Final Story Points to 4.