1065553 - Document deployment-specific links & instructions on Mana

Status: RESOLVED DUPLICATE of bug 1165259

(Tree Management :: Treeherder: Infrastructure, defect, P4)

Description

[Ed Morley [:emorley]]

It would be good to document:
* The hostnames/IPs for dev/staging/prod
* Explanations of each host (eg <foo> does log parser, <bar> serves xyz)
* How to request access for each (ie IT for prod, glob for dev/staging on AWS) as well as VPN access for Chief.
* The .ssh/config examples for them (eg for the AWS hosts: https://intranet.mozilla.org/Auto-tools/Systems/AWS/Creation)
* How to access New Relic
* Chief deploy details for prod
* Deploy instructions for dev/staging
* Locations for error logs for ingestion/apache/... and which hosts to check for certain issues, in case treeherder-devs are not around and there is a problem in production.

Comment 1

[Ed Morley [:emorley]]

Moving this to infra (vs the docs component), since I see this bug as more prod/stage/dev specific (vs treeherder architecture for bug 1068799) and the content will likely need to be on Mana or similar given IP addresses, vs bug 1068799 which will be generic and likely in Readthedocs.

Comment 2

[Ed Morley [:emorley]]

https://mana.mozilla.org/wiki/display/websites/treeherder.mozilla.org now has some of this - missing is:

> * The hostnames/IPs for dev
> * How to request access for each (ie IT for prod, glob for dev/staging on
> AWS) as well as VPN access for Chief.
> * The .ssh/config examples for them (eg for the AWS hosts:
> https://intranet.mozilla.org/Auto-tools/Systems/AWS/Creation)
> * How to access New Relic
> * Chief deploy details for prod/stage
> * Deploy instructions for dev
> * Locations for error logs for ingestion/apache/... and which hosts to check
> for certain issues, in case treeherder-devs are not around and there is a
> problem in production.

Comment 3

[Ed Morley [:emorley]]

Also "how to manually purge a queue from the cli", eg:
celery -A treeherder worker -Q default --purge

Comment 4

[Ed Morley [:emorley]]

To clear memcached:
from django.core.cache import cache; cache.clear()

Comment 5

[Ed Morley [:emorley]]

The few bits of this that are still relevant will be done in bug 1165259.

> * The hostnames/IPs for dev/staging/prod

Will link to the Heroku pipeline from Read the docs

> * Explanations of each host (eg <foo> does log parser, <bar> serves xyz)

wontfix - see bug 1068799 comment 2.

> * How to request access for each (ie IT for prod, glob for dev/staging on
> AWS) as well as VPN access for Chief.

Heroku access is all that is required

> * The .ssh/config examples for them (eg for the AWS hosts:
> https://intranet.mozilla.org/Auto-tools/Systems/AWS/Creation)

No longer applicable once we're using Heroku.

> * How to access New Relic

Can link to New Relic from read the docs.

> * Chief deploy details for prod
> * Deploy instructions for dev/staging

Can mention staging deploys from master, and process for deploying prod (likely pushing to a prod branch).

> * Locations for error logs for ingestion/apache/... and which hosts to check
> for certain issues, in case treeherder-devs are not around and there is a
> problem in production.

Will just link to papertrail

> Also "how to manually purge a queue from the cli", eg:
> celery -A treeherder worker -Q default --purge

Can link to the rabbitmq admin page on cloudamqp

> To clear memcached:
> from django.core.cache import cache; cache.clear()

Can link to the memcachier page, which has a clear cache button.