The Delivery Team maintains a support rotation where an engineer is dedicated to providing support to our internal teams for a term of 1 week.
The support rotation schedule is captured in OpsGenie. The support rotation consists of an Engineer On-Call (EOC) and a technical manager (TM).
On-Call duties transfer automatically Mondays at 10am Central Time according to the schedule within OpsGenie.
The Delivery team aims to provide timely responses that are proportional to the capacity of the team.
The Engineer On-Call (EOC) is responsible for a 24 hour x 7 period shift where an initial response or acknowledgement is expected within 10 minutes.
After 10 minutes, the Technical Manager is paged and expected to provide an initial response or acknowledgement within 5 minutes.
This is designed to ensure that an initial response is produced within 20 minutes of the page.
Note: ⚠️ An OpsGenie page is considered an emergency. It may wake up the EOC in the middle of the night. It’s expected to be used with the utmost discretion.
The Delivery team aims to facilitate and uphold the SLAs maintained by the Customer Support team. The Delivery team will triage and provide an initial response within 24 hours of receiving a support request or request for help.
The engineer that is starting their support rotation.
The engineer that is ending their support rotation.
The engineer that is currently support rotation.
The manager of the Delivery team exists as the terminal escalation for the support rotation.
At the start of their shift, the Incoming Engineer On-Call (IEOC) will prepare to take over support rotation duties from the Outgoing Engineer On-Call.
This document is designed to help the Incoming Engineer On-Call get up to speed with any open issues that require their attention during their shift. Additionally, this is designed to provide a mechanism for the Outgoing Engineer On-Call to offload support duties in order to transition back into project work.
This can be done in the OpsGenie interface by supplying your cellphone number or by downloading the OpsGenie app. The method you choose to be contacted is up to you as long as your response time follows the Response Time guidelines
The Delivery Team OpsGenie Dashboard is: https://sourcegraph.app.opsgenie.com/teams/dashboard/9fac6e04-e826-4e3d-a1bb-1790e833bba5/main
OpsGenie has documentation that describes how to setup Notification Preferences.
The following Slack channels should be monitored during your shift. There is no expectation to actively monitor these channels 24/7 during your support week, but you should serve as the designated point of contact during your working hours.
|Channel Name||Description||What to look for|
|#delivery||Public channel where Sourcegraph teammates can ask questions, provide information, and escalate support issues.||Questions from other teams, request for help, and other Delivery-related information.|
|#incidents||Company-wide incident feed.||Incidents that may affect on-prem/managed instance deployments.|
|#prod-eng-announcements||Product Engineering Announcements.||Monitor for release-related items or larger feature roll-outs that may affect on-prem / managed instance deployments.|
|#ask-prod-eng||Company-wide forum to discuss Product Engineering topics.||Generally questions are tagged with “[Delivery]”|
|#buildkite-main||Information about BuildKite CI.||The build checker will occasionally lock the main branch when the main build fails to build. When serving as a release captain it may be helpful to know the current state of the build.|
You will be automatically added to the
@delivery-support Slack group by OpsGenie. This handle is used to call attention to
important topics in Slack that may or may not include incidents or P1 customer support issues.
While an immediate response is not necessarily required, use your best judgement to determine an appropriate response.
This is section is not designed to be prescriptive. You should use your best judgement on how to handle your support rotation duties. This is by no means an exhaustive list but some notes of things to check while on-duty.
Delivery engineers who are not support rotation do not need to respond to questions, instead they should defer to the Engineer On-Call (EOC). This is intended to promote knowledge transfer from subject matter experts (SMEs) to the rest of the Delivery team.
While the EOC may not provide the final resolution, the expectation is to provide an initial response, triage, and ensure that issues are captured in the Delivery Github Project.
Anything we are not doing today or not documented in our handbook is most likely a feature request.
If there is a known gap in functionality, or extensive work necessary to service a request, (e.g., a feature request), escalate to the manager for prioritization.
The EOC is encouraged to use the handbook, docs site, other resources, or consulting the larger Delivery team to provide as much context as possible when during triage and providing an initial response.
You are not expected to monitor slack 24/7. During your working hours periodically check for messages in the Slack Channels listed above.
More urgent Slack messages are usually tagged with
OpsGenie is configured with team-based escalations. OpsGenie will page the scheduled Engineer On-Call. Pages from OpsGenie should be considered a priority and follow the guidelines for Response Times.
OpsGenie is configured to page the EOC when:
- A Sourcegraph teammate engages OpsGenie to page the EOC.
Keep notes about any incidents, alerts, and priority support issues during your shift. They’ll come in handy when you transition your duties to the next Incoming Engineer On-Call.
Having a formal end to your support rotation shift is intended to provide a way to debrief and promote a healthy relationship with support rotation duties to reduce burnout.
While not prescriptive, you should take measures to reduce disruptions so you may focus on your project work.
Reduce disruptions while you are no longer on-duty!
You will be automatically unsubscribed from the
@delivery-support Slack handle.
As the Outgoing Engineer On-Call, your responsibility to ensure the Delivery On-Call Hand-off Notes are up-to-date and ready for the Incoming Engineer On-Call by 10am Central Time on Monday.
If you responded to an incident or had an eventful support rotation, consider taking a day off to recover and prevent burnout. Sourcegraph has an unlimited PTO policy.