Data and Analytics Architect: Data and Analytics vision and plan signed off by executive team NOT STARTED
Data and Analytics Architect: Create a user journey funnel NOT STARTED
CMO: Educational Email campaign series to educate users on full solution capabilities and how to get started. TODO # of campaign conversions => Issue board anniversary awareness campaign on blog and social. Initiated CI webcast series.
Production: Scalable and HA setup for cache and background jobs storage.
Persistence of cache (Redis) and background job storage (Sidekiq) split and
HA is set up for persistent storage. => Done.
Production: Robust backups with automatic periodic restores. Counter on
dashboards.gitlab.com shows days since last automatic restore of database and file
system. => Backups are tested but restores are not automatic, and counter is missing.
Production: Multi-canary deployments enabled. PoC of two canaries via kubernetes with Distribution team's Helm charts. => One canary available, not via kubernetes.
Production: Enable GitLab.com search using ElasticSearch. Provide a cluster that is 99% available for > 2 months in Q3. => Infrastructure available but application requires further work.
Database: Run pgbouncer from Omnibus. Shipped. => Done
Database: Use Omnibus provided software for HA / failover instead of using a GitLab.com specific setup. Shipped and In Use. => Not yet in use.
VP Eng: Enable High Availability
Platform: Enable graceful degradation when file servers are down. => Circuitbreakers built but not yet fully deployed.
Platform: Support multiple Redis clusters for persistent and cache stores. => Done.
Platform: Use Geo DR to move between clouds and Area Zones => 10% done, begin testing with Geo testbed, pushing to Q4.
Support: Improve GitLab provided debugging tools. We'll log all times we need terminal access and create an issue for each to improve GitLab tools. => 50% done, see https://gitlab.com/gitlab-com/support/issues/754
Prometheus: Reach parity with Prometheus metrics for Unicorn, Sidekiq, and gitlab-shell. Deprecate InfluxDB. => 30% done with Prometheus metrics for Unicorn. Sidekiq and gitlab-shell remain for Q4.
Gitaly: Gitaly service active on file-servers. => Done.
Gitaly: Roll out Gitaly migrations. 24 additional endpoints migrated to Gitaly and in acceptance testing. => Done.
Gitaly: Reduce “idea to production” time of migrations. 80% of all migrations started in Q3 are enabled on GitLab.com within two GitLab releases. => Done by changing process.
Database: Reduce the p99 of SQL timings across the board to 200 ms (100-200 ms less than what we have now). => Significant progress, but not all p99 < 200 ms. See gitlab-ce#34535 for more context.
Production: Provide reliable internal & external
baseline monitoring of overall service health. Implement baseline end to end
monitoring for GitLab.com (api/web/git(ssh/https)) and define
based on this baseline. => Baseline not defined.
Production: Solve performance issues. Implement CDN for GitLab.com. => CDN was implemented but had to be rolled back.
VP Eng: Lower latency in application
Discussion: Solve performance issues. Reduce p95 of discussion-related actions with over 10 hits/day to < 1 s. Reduce p99 to < 3 s. => 25% done. 10% (~150 of 1476 endpoints) still exceed thresholds. http://stats.gitlab.com/1902794 is at 1.2 s.
Platform: Solve performance issues. Reduce p95 of platform-related actions with over 10 hits/day to < 1 s. Reduce p99 to < 3 s. 25% done. 10% (~150 of 1476 endpoints) still exceed thresholds.
Discussion: Merge requests get merged 100% without ever getting into stuck locked state => 20% done. Improvements made in https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/13207, but still occurring. We need to measure rate.
Platform: Project authorizations fast (< 1 s) and consistent (requires no manual refreshes). DONE
Platform: Project imports and forks complete 100% without ever getting stuck => 10% done. Import issues need to be broken out separately.
Platform: Namespace and project renames work 100% of the time => 30% done. Requires hashed storage support
Platform: Repository cache state eventually consistent within minutes (no manual expiration needed) => In progress with https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/14980
Security: Improve security practices through people and processes. Backlog of security issues reduced by 50% (currently 148 issues in gitlab-ce, 9 in gitlab-ee, 69 on infrastructure). => Not done, backlog is larger.
CEO: Increased usage of idea to production. 100% growth
CMO: More contributors from the wider community each month. Unique contributors grow 10% QoQ => Missed.
CMO: Initiate AR engagement with key analysts and achieve Leader in the Forrester Wave CI research report. => Leader in Forrester CI Wave. Account engagement with Forrester/Gartner. Analyst day and multiple briefings.