Some alerts are urgent enough to warrant immediate attention, such as parts of the site becoming unavailable or large quantities of error pages being served. We use PagerDuty to escalate to these people (in order):
- Primary Engineer
- Secondary Engineer
- Programme Team member (might not be technical)
It is the responsibility of the people above to make sure their details are up to date in PagerDuty and that they’re correctly scheduled in. You can add your rota to Google calendar.
When an alert that triggers PagerDuty goes off, someone on the escalation schedule must acknowledge them, otherwise they will be escalated further. NB, Technical 2nd Line shadowers are not required to be on PagerDuty.
Creating a schedule override
There are times where developers need to make rota swaps or cover Technical 2nd Line last minute. See the Pagerduty documentation on how to schedule an override.
Every week we test PagerDuty to make sure it can phone to alert us to any issues. This happens every Wednesday morning at 10am UTC.
Cron creates a file on disk. Icinga raises a critical alert when that file exists with an Icinga contact group set so that PagerDuty is notified. After a short amount of time, cron will remove the file to resolve the alert. The code that does this is in Puppet.
You don’t need to take any action for this alert. The primary in-office Technical 2nd Line developer should escalate the call to the secondary who should escalate it to “escalations” to ensure that phone redirection is working. The person on “escalations” will resolve the PagerDuty alert to prevent anyone else being phoned.
Triggering the drill manually
To trigger the drill manually, follow these steps:
SSH onto the monitoring box:
$ gds govuk connect -e production ssh aws/monitoring
Generate the file that the Icinga monitor looks for which triggers the alert:
$ sudo touch /var/run/pagerduty_drill
Delete the file once the test drill is triggered:
sudo rm /var/run/pagerduty_drill
Add your Pagerduty rota to Google calendar
You can sync your Pagerduty with Google Calendar so you can see your Technical 2nd Line and on-call shifts in one place.
Go to https://governmentdigitalservice.pagerduty.com/my-on-call/month
On the left side, click “Export Calendar”, right click on “WebCal Feed” and copy the link
In Google Calendar, under “other calendars” click the plus symbol and add “From URL”