Update New Relic redis monitoring configuration

[rioug]

Description

This came up as part of investigating infinite loading bug : openfoodfoundation/openfoodnetwork#11752 (comment)

New Relic provides monitoring for redis, however it's configured with the wrong port. We need to be monitoring redis on port 6380. Redis monitoring is accessible in New Relic via Infrastructure -> Third party services -> Redis

• production config : https://github.com/openfoodfoundation/openfoodnetwork/blob/6240f37adfc3993b1ce18decb2024a00968e77a7/config/environments/production.rb#L46
• staging config : https://github.com/openfoodfoundation/openfoodnetwork/blob/6240f37adfc3993b1ce18decb2024a00968e77a7/config/environments/staging.rb#L46

Server to fix :

• au_prod, we need to fix all third party monitoring see : New relic, fix third party monitoring on au_prod #923
• au_staging
• ca_prod
• hu_prod
• ie_prod
• in_prod New Relic not setup
• nz_prod
• uk_prod
• uk_staging
• uk_staging2?
• us_prod
• es_prod
• fr_prod
• fr_staging
• de_prod
• be_prod

TODO

Fix ansible script so the correct configuration is used next time we provision a server

Fix

As an ofn-admin, ssh to the server:

• cd /etc/newrelic-infra/integrations.d
• update the PORT entries to 6380 in redis-config.yml
• restart new relic service : systemctl restart newrelic-infra

Expected Behavior

Relic monitoring reports redis activity :

Actual Behaviour

Relic monitoring shows a flat graph :

[dacook]

I wondered if we can set this up with Ansible?

Even better, it should probably be an .env variable, which both openfoodnetwork and newrelic could refer to.

[rioug]

All instance be au_prod (new relic not setup for third party monitoring ) and in_prod (new relic not setup) have been manually updated and are now reporting in new relic.

Leaving the ticket open till we fix the ansible script