1395210 - Request for simple ping monitoring of seven websites Bugzilla

Status: RESOLVED FIXED

(Infrastructure & Operations :: MOC: Service Requests, task)

Description

[Daniel Hartnell [:danielh]]

Hi,

I am moving these websites to AWS via Nubis. Can we ensure that we have a ping monitor for each site to keep track of availability?

static.mozilla.com
website-archive.mozilla.org
planet.bugzilla.org
planet.firefox.com
planet.mozilla.de
planet.mozilla.org
planet.webmademovies.org	

Let me know if you have any questions! Thanks.

Comment 1

[Keegan Ferrando [:fauweh]]

Also, from a comment in CAB, please make these HTTP checks for status versus ICMP checks. Nubis applications are commonly behind ELBs and they often are not configured respond to ICMP.

Comment 2

[Keegan Ferrando [:fauweh]]

Daniel, I'll add these in NR Synthetics. 

Can you tell me if we should be checking HTTP, HTTPS, or both? Also, are there any string searches we should do upon a 200 response or just validate we end up with a 200 OK?

Thanks!

Comment 3

[Daniel Hartnell [:danielh]]

Hey Keegan,

I see no reason to go beyond checking for a 200 response. Depending on how Synthetics deals with redirects, you might have to account for that. These sites will redirect HTTP to HTTPS connections and they may return a 301 prior to the 200. You can test with tlscanary.mozilla.org if you want. All these sites should function similarly to that (also hosted under Nubis Haul).

Comment 4

[Daniel Hartnell [:danielh]]

I spoke with Sal regarding static.mozilla.com. The current production site displays a generic Apache welcome page and this will not be carried over to the Nubis deployment. This means that unless I provide the path to a specific file we are serving, you will always get a 403 response code. I don't want to provide the path to a specific file because all of these are managed through a GitHub repository and a change could result in a broken monitor. If your monitor can succeed with a 403 response on the index, that's fine with me. If you need a 200, I have setup a health endpoint that you can request at /health. It still needs some work but that should be a place once I make the production switch this evening (5-6pm pacific). Feel free to reach out to me with any questions.

Comment 5

[Keegan Ferrando [:fauweh]]

Ok, I reviewed with Daniel after the migration today. All sites listed in the description are now added in New Relic and have PagerDuty alerts enabled. 

As the static.m.c was rolled back, it current returns a 200 OK and is working. When this is migrated to Nubis, we will need a a URI that returns a 200 OK to monitor.