roam/daily/2021-12-09.org
2021-12-13 17:29:02 -05:00

4.5 KiB

2021-12-09

CREASE

Session with Steve Eisner.

Reviewing the notes on what we wanted to achieve in our last session, which was cut short.

Steve's current queue is a bit sparse, so we're looking through the Top ASE Customer Requests and Upcoming Product Team Backlogs - CSLead SLAs sections of the ASE structure board.

Etsy imports

Requires less babysitting now. Better than before, but not quite all the way there.

Campaign messages

Messages attached to a campaign and subsequently removed aren't fully untethered from campaigns, causing them to not show up in drafts. Steve has Jupyter notebooks for correcting the issue.

Export broadcast quickstats in bulk

Should be much less relevant now that clicks and opens are in the account export.

Create a way for customers to automatically merge lists

Steve would discard some data, including custom fields, to simplify the manual process.

Ability to bulk tag over 50,000 subscribers

Steve wonders whether this could be worked around by using update-on-import (does that have a row limit? 20k?).

View duplicate subscribers across lists

Steve wonders if the customer is aiming to lower their bill?

Remove and edit previous contact addresses from account

I suggested that it is probably best to add a flag to retire it / keep it from being displayed to the user.

Bulk delete inactive subscribers across many lists

Steve believes this makes sense to keep this in ASE, warns that in the cases of mistakes things have to be restored, which can be difficult and miss things (Admin restore deleted controller). The subscriber ID should be restored now, but are the mappings?

Customer export file size limited to 100MB (now 200MB)

Due to a cap in file upload sizes in Tornado.

As an AWeber customer, I would like a report of tags applied on a given day

Steve managed this once using Athena queries, doesn't recall requests from any other customers.

Broadcast Automations are not Applying/Removing Tags

Steve looked into this and was able to replicate the issue on one of his own subscribers.

Restoring a subscriber doesn't restore the recipient record (remains in a deleted state).

Average Response Time alert - Subscriber Proxy

Pagerduty alert.

Seems related to bulk tag job retrieval (JobHandler.GET).

This aligns with requests in the bulk tagging logs.

  {
    "asctime": "2021-12-09T18:20:45.538764+0000",
    "correlation_id": "-",
    "levelname": "INFO",
    "message": "10.51.23.1 - - [09/Dec/2021:18:20:44 +0000] \"GET /jobs?account_id=1044937 HTTP/1.1\" 200 - \"-\" \"subscriberproxy/1.2.5\" (secs:1.281)",
    "module": "helpers",
    "name": "tornado.access",
    "service": "bulk-tagging",
    "exc_info": null
  }

This account in particular is a test account with a large number of jobs, and it appears to be regularly requested. It looks like bulk-tagging isn't successfully applying limits and ends up taking a while to collect a large number of jobs.

  GET http://bulk-tagging.service.production.consul/jobs?account=1044937&limit=3
4530