Context

• In this case the resume was not imported.

• Workflow (Lucid Lin)

  

Checks

1. The amount of connections, Memory Ram and disk increase without error. (Link)

   

   • Alert by RAM was raised, Post finish take 30 minutes to return connections to normal behave, some connections have the status [idle](https://www.enterprisedb.com/postgres-tutorials/how-monitor-postgresql-connections#:~:text=idle%3A This indicates that the,input from the end user.), but they were solved automatically.
     • Screenshot: PostgreSQL return the normality
     • Screenshot: connections in idle estatus
     • Queries used for check the status

2. Python API MySQL (link)

   

3. Heroku Python API could be affected (Link)

• Replica server looks normal.(Link)

Amount of 5xx error(Link)

we can process >600 per minute (Link)

4. Heroku Integrations(Link)

Recomendation

1. Increase the WEB_CONCURRENCY in python api to the same number of the replica app. 60 → 84

2. Evaluate if it is convenient to change the type of dyno for the worker the python api, the memory is very close to the top.