Infrastructure/Oncall

Fedora Infrastructure Oncall

What is this Oncall thing anyhow?

The Fedora Infrastructure oncall person is someone who has been designated by our team to handle incoming requests for work and triage them for the current week.

How can I find out who is oncall?

On irc you can use the 'oncall' alias. If you are in a channel with the Fedora zodbot user, you can do '.oncall' if you are not, you can '/msg zodbot oncall'

Why is there one oncall person?

Fedora Infrastructure is composed of a small group of full time Red Hat Employees and community members. When someone communicates directly to a team member they interrupt them from whatever they were working on. Your request could well be less important than the task they were working on and even if they stop to tell you that it will take a while to get back into the task they were doing. Many studies have shown it could take 20-30minutes to regain concentration after being interrupted. If everyone talks to the same oncall person they can triage and/or handle your request without bothering the rest of the team that may be working on other tasks. Additionally, team members not on call can plan longer tasks that require concentration for the times they are not on call.

I was directed to this page, why?

If you tried to contact a team member directly about a work item and were directed here, please ask the oncall person (or if they are not available, file a ticket).

Most likely it's because you bypassed the oncall person and interrupted another team member Please don't do that.

Additionally, please avoid pinging team members for non urgent/notice related matters. (ie, don't file a ticket or PR, then ping a team member right after that noting that you did so, they will see it in their email)

No naked pings

Please don't ping anyone without information to be able to tell what it was about. See: No naked pings

But the oncall person isn't answering me

If there's no reply from the oncall person, please file a ticket at https://pagure.io/fedora-infrastructure/new_issue. This will be noticed by other team members as they have time to look and/or the oncall person can triage it after they are back.

What things can/does the oncall person handle?

The oncall person should answer pings for the team and determine if the request is urgent enough to interrupt someone for, or if they can process the request they can do so, or they can direct that a ticket be created.

The oncall person can triage tickets (move from the 'needs review' state to another state) as their time permits.

The oncall person can answer alerts / notifications as their availability permits. Ideally they would ack or fix alerts before they go to pagers on the second alert.

How to set the On-call message

The oncall message set in zodbot should state who is oncall, how they can be reached, and what their normal working hours and days are on the UTC clock. An example could be set like the following

zodbot: alias add oncall "echo mattdm is oncall. Please file a ticket if you don't hear from me ( https://pagure.io/fedora-infrastructure/issues ) My current hours are 1300 UTC to 2100 UTC Monday through Friday"

If there is a time when no one can be on call, please put up the following:

zodbot: alias add oncall "echo No-one is currently oncall. Do not expect pings to be seen or remembered. Open a ticket at https://pagure.io/fedora-infrastructure/issues "