There's definitely a very big challenge in observability of large systems. It's become impossible to know what everyone is doing and experiencing at a given point in time. Part of that is sort of the engagement metrics we talked about earlier, how do we actually track interactions and what people are actually doing? And other parts of it could be how do we have safety nets through user reporting.
Jack Li explains how his production engineering team rolled out a new incident review process, how they’ve made the case for investing in reliability, and specific tools his team has built to improve reliability.
—
Discussion points:
- (1:25) How Jack became interested in reliability
- (3:24) Where the Instagram Reels team fits into the broader organization
- (4:05) What Jack’s team focuses on
- (4:55) The role of production engineering at Instagram versus Shopify
- (8:32) The essence of DevOps
- (10:44) Pros and cons of having product-focused teams
- (13:35) How Jack’s team defines and tracks quality
- (15:46) Signals the team monitors outside of systems
- (18:10) Revamping Instagram Reel’s incident management process
- (19:46) Making the case for improving the incident review process
- (28:10) How their incident review process works
- (31:55) The roles involved in an incident review
- (33:40) The value of having incident reviews
- (35:55) Why leaders should be part of incident reviews
- (38:34) Why Jack’s team builds tools for driving reliability goals
- (40:06) The types of tools Jack’s team focuses on
- (43:09) What a merge queue is and why it was built at Shopify
- (51:20) Using a Slack bot for ‘failed build’ alerts
- (52:32) When a company should consider implementing a merge queue
—
Mentions and links:
Follow Jack on LinkedIn
Jack’s article from his time on Shopify about their Merge Queue
Jack’s talk on Shopify’s Merge Queue at GitHub Universe 2019