Wikimedia Technology/Annual Plans/FY2019/TEC3: Deployment Pipeline/Goals
Program Goals and Status for FY18/19
edit- Goal Owner: Greg Grossmeier
- Program Goals for FY18/19: We will streamline and integrate the delivery of services, by building a new production platform for integrated development, testing, deployment and hosting of applications. Wikimedia developers experience a tooling parity between our Continuous Integration (CI) and production environments which enables them to release code more frequently by continuously reducing risk.
- Annual Plan: TEC3 Deployment Pipeline
- Primary Goal is Knowledge as a Service: Evolve our systems and structures
- Tech Goal: Sustaining
Outcome 1 / Output 1.1
editContinuous Integration is unified with production tooling and developer feedback is faster
- Convert current CI builds to use the new tooling (Blubber).
Dependencies on: SRE team
Goal(s)
edit- Move verify stage from Minikube to CI k8s namespace in production context
Status
editNote: July 2018
- In progress
Note: August 10, 2018
- In progress Discussed that work on a patch is still ongoing, need to refactor the pipeline job to the new namespace. This will be a change to the existing service but will need to be refactored when we get to the shared library.
Note: September 14, 2018
- This is now Done!
Outcome 1 / Output 1.2
editContinuous Integration is unified with production tooling and developer feedback is faster
- Setup test execution time profiling with a report, make a prioritized list of improvements to how tests are run.
Goal
edit- Formalize the collection of CI infrastructure and tooling metrics - task T205923
Status
editNote: October 2, 2018
- This is now In progress
Note: November 7, 2018
- In progress
- dduvall gave a presentation Monday looking at CI performance percentiles
- Work continues on automating the collection of these metrics.
Note: December 6, 2018
- This goal is Partially done but we need to expose the interface of the metrics that we're collecting.
Outcome 2 / Output 2.3
editDeployers have a better assessment of risk with each deploy
- Improve our incident response, post-mortem, and follow-up management tooling.
Goal
edit- Develop set of metrics to assess incident reports/post mortems.
Status
editNote: October 2, 2018
- This work has not yet been started at this time To do
Note: November 7, 2018
- This is now In progress with Zeljko's analysis of the past year's worth of incident reports.
Note: December 6, 2018
- We can now determine how associated commits, repos, etc are connected to the incident reports and consider this goal Done. There is more work that we can do to further refine the metrics.
Outcome 3 / Output 3.1
editDeployments happen through percentage based stages (eg: canaries, 10%, 100%)
- Migration of services currently on our "shared service cluster" into Kubernetes (k8s) deployments with staged rollout
Primary teams: Service Operations, Release Engineering
Goal(s)
edit- Adopt more services into Deployment pipeline
- Migrate graphoid to the Deployment pipeline Postponed
- Deploy zotero v2 to the Deployment pipeline Done
- Deploy blubberoid Done
- Reprise the work on the logging infrastructure task T207200
Status
editNote: October 2, 2018
- This is now Done
Note: November 7, 2018
- Deploy zotero v2 to the Deployment pipeline Done
- Currently living in k8s staging
- Plan to go live next week
- Deploy blubberoid Done
- liw working on changes to internal data structuring as a prerequisite to creating OpenAPI spec required for pipeline — on track.
- Deploy zotero v2 to the Deployment pipeline Done
Note: December 6, 2018
- Zotero is Done, Graphoid will be recommended for stewardship review Postponed, Blubber will be Done in the next week. Reprise the work on the logging infrastructure is still Done.
Outcome 1 / Output 1.2 (RelEng)
editContinuous Integration is unified with production tooling and developer feedback is faster
- Setup test execution time profiling with a report, make a prioritized list of improvements to how tests are run.
Goal
edit- Instrument Quibble for data collection
- Create a graph where time is spent and make a prioritized list for improvements.
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
Note: February 5,2019
- These task has been documented in Phab during last week's all hands meetings.
Note: March 12, 2019
- This goal is currently in danger of finishing up this quarter and will be part of Q4 goals Stalled
Outcome 2 / Output 2.1 (RelEng)
editDeployers have a better assessment of risk with each deploy
- Create a deployments report with metrics from the Code Health Group.
Goal(s)
edit- Select and integrate a code health metric solution into our tooling.
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
Note: February 2019
- Discussed that this is contingent on some other program work in TEC13, but we should be able to get fully started on it soon.
Note: March 12, 2019
- We've selected sonarcube to be our metric solution, but we need to create and finalize the integration for it (with sonarcloud) for self hosting and get it integrated into CI (still In progress and moved into Q4 work.
Outcome 3 / Output 3.1 (SRE / ServiceOps & RelEng)
editDeployments happen through percentage based stages (eg: canaries, 10%, 100%)
- Migration of services currently on our "shared service cluster" into Kubernetes (k8s) deployments with staged rollout
Primary teams: Service Operations, Release Engineering, Core Platform
Goal(s)
editAdopt more services in the pipeline
- cxserver, ORES (partially), citoid, changeprop, cpjobqueue (stretch)
- Deploy eventgate
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
Note: February 5, 2019
- Discussed that there is one service that is currently (event gate) going through the pipeline...still In progress
Note: March 2019
- cxserver: all prerequsites are ready, just need to be deployed; ores: still In progress and blockers identified; citoid Done; changeprop is still In progress; cpjobqueue is Postponed to Q4 and eventgate has been deployed and is Done
Note: April 8, 2019
- cxserver is now Done
Outcome 4 / Output 4.1 (SRE / ServiceOps)
editDevelopers are able to create services that achieve production level standards with minimal overhead
- Developers get a service creating experience that is on par with production level standards with regard to logging, monitoring, security and configuration
Goal(s)
edit- Evaluate helm charts management solutions
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
Note: February 5, 2019
- Discussed this at All Hands, but we're working with SRE on it, still In progress
To do March 2019
- Discussed...
Outcome 5 / Output 5.1 (SRE / ServiceOps)
editServices and the deployment pipeline are hosted on production-level infrastructure
- Adequately maintain the service infrastructure according to production standards. Upgrade the platform to new upstream versions to benefit from bug fixes and new features.
Goal(s)
edit- Aim for a better resilient, scalable, easier to manage and upgrade Kubernetes cluster service
- Upgrade cluster components to a newer version
- Improve docker registry architecture
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
To do February 2019
- Discussed...
To do March 2019
- Discussed...
Outcome 6 / Output 6.1 (SRE / ServiceOps)
editDevelopers and deployers are aware of the platform, its benefits and how to make use of it
- Create a developer portal for the Deployment Pipeline platform with documentation and instructions
Goal(s)
edit- Create a developer portal with Deployment Pipeline documentation
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
To do February 2019
- Discussed...
To do March 2019
- Discussed...
Outcome 6 / Output 6.2 (SRE / ServiceOps)
editDevelopers and deployers are aware of the platform, its benefits and how to make use of it
- Promote the platform's adoption
Goal(s)
edit- Conduct Promotion and Training events for Wikimedia developers
Status
editNote: January 10, 2019
- Discussed that as we've just gotten back from our vacations, this work is ramping up and is In progress
To do February 2019
- Discussed...
To do March 2019
- Discussed...
Outcome 1 / Output 1.2 (RelEng)
editContinuous Integration is unified with production tooling and developer feedback is faster
- Setup test execution time profiling with a report, make a prioritized list of improvements to how tests are run.
Goal
edit- Instrument Quibble for data collection
- Create a graph where time is spent and make a prioritized list for improvements.
- Prepare the Deployment Pipeline for changes to our CI tooling.
Status
editNote: April 8, 2019
- This is now In progress, but the instrumenting Blocked right now.
Note: May 7, 2019
- This is still Blocked, as we are waiting on other teams.
Note: June 4, 2019
- This is still Blocked, as we are waiting our team to get done with other more urgent work, will probably go into next FY.
Outcome 3 / Output 3.1 (RelEng + SRE)
editDeployments happen through percentage based stages (eg: canaries, 10%, 100%)
- Migration of services currently on our "shared service cluster" into Kubernetes deployments with staged rollout
Goal(s)
edit- Create a .pipeline/config.yaml standard to give users more control over how their tests are run in the pipeline and allow the easy saving of artifacts at pipeline completion. (RelEng)
- Migration of more services to the pipeline (RelEng + SRE) - task T212801:
- Wikidata Termbox SSR
- Kask for Session Storage Service
- cpjobqueue (stretch)
- ORES (stretch)
Status
editNote: April 8, 2019
- All goals are In progress
Note: May 7, 2019
- All goals are In progress
Note: June 4, 2019
- All goals are In progress, termbox is now producing images, but not yet in production; kask is also Partially done. Stretch goals will probably move to next FY. Pipeline config work is Done
Note: June 13, 2019
- Migration should be done this quarter, but the overall program will remain In progress into next FY, stretch goals will be done next FY. (SRE update)
Outcome 5 / Output 5.1 (SRE)
editServices and the deployment pipeline are hosted on production-level infrastructure
- Adequately maintain the service infrastructure according to production standards. Upgrade the platform to new upstream versions to benefit from bug fixes and new features.
Goal(s)
edit- Upgrade the infrastructure to recent/current software versions
- Add dedicated security sensitive nodes to the Kubernetes clusters
- Stretch: Implementation of a Helm chart management solution
Status
editNote: June 13, 2019
- Upgrade is still In progress and will be done by end of quarter. The addition of the dedicated nodes is Done and the stretch goal was just recently started and will be done by the end of the quarter too.
Outcome 6 / Output 6.1 + 6.2 (SRE)
editDevelopers and deployers are aware of the platform, its benefits and how to make use of it
- Create a developer portal for the Deployment Pipeline platform with documentation and instructions
- Promote the platform's adoption
Goal(s)
edit- Conduct at least N trainings for new pipeline users
- Increase documentation quality
Status
editNote: June 13, 2019
- Trainings have been on-going and still In progress
- Documentation quality has been going slow, and will probably go into next quarter.