# 2021-05-14 Price range not getting updated postmortem

Date: 2021-05-14

Summary: Price range was updating erratically on production environment

Impact:

  • Users could not view the price range given a search criteria
  • Users were under the false impression that price range was still loading, even though it had failed to update.

Root causes: AWS Appsync was down globally for an hour.

Resolution: Monitor AWS updates about App Sync.

Detection: Manual

Action items:

Action item Type
Verify Price range in synthetic monitoring process
Update cloudwatch alarms process

# Lessons Learned

# What went well

  • Ali identified the issue on production early on.
  • We spent minimum time to understand what actually went wrong.
  • Our acceptance test suite gave us the confidence that code wasn't broken.

# What went wrong

  • Synthetic monitoring did not catch this.

# Where we got lucky

  • AWS is quite prompt at resolving issues with their infrastructure.
  • Not a lot of active users

# Timeline

2021-05-14 (all times UTC, unless explicity mentioned)

  • 16:05 Ali notifies price range is loading intermittently on prod.
  • 16:06 Harsh verifies and confirms he sees the same consistently on prod. The price range text was stuck at Loading... and the browser console showed 500 errors.
  • 16:06 Harsh navigates to graphql lambda logs and doesn't find any. This confirms our service wasn't getting hit at all. Potential cause could be CORS errors from the API.
  • 16:07 Harsh tries again on staging and it works.
  • 16:07 Harsh tries again on production and it works.
  • 16:08 Harsh navigates to AWS notification console and finds a notification from AWS: Event- Appsync API issue
  • 16:08 Harsh notifies this to the team and waits for further notifications form AWS.
  • 16:09 Harsh randomly verifies prod again and finds it's still working. The issue is still open with AWS.
  • 20:30 (PDT) - AWS confirms the issue is resolved from their end