Wikimedia Cloud Services team/goals/2023-24

Q3/Q4 (Jan 2024 to June 2024)[edit]

Phabricator board

Goals[edit]

Incomplete goals from Q1+Q2[edit]

Toolforge
- phab:T332955 [maintain-dbusers] Generate prometheus metrics
Cloud VPS
- phab:T328502 Move WMCS off of Icinga and introduce alertmanager
- phab:T341060 openstack eqiad1: introduce cloud-private and cloudlb
- phab:T309789 [ceph] Upgrade hosts to bullseye
- phab:T306820 [ceph] Upgrade to v16
- phab:T297083 [ceph] Getting rack level HA
Data Services
- phab:T291782 Migrate largest ToolsDB users to Trove

New goals for Q3/Q4[edit]

Toolforge
- phab:T314664 Toolforge: Decommission the Grid Engine infrastructure
- phab:T194332 Some form of push-to-deploy on Toolforge
- phab:T311897 Migrate Toolforge off of Debian Buster
- phab:T336668 [harbor] Create backups and/or replication
- phab:T350687 [harbor] Move harbor data to object storage service
- phab:T356301 [harbor] Deploy with Helm
Cloud VPS
- phab:T351450 Migrate Cloud VPS puppet infrastructure to Puppet 7
- phab:T356287 Upgrade to openstack version Bobcat
- phab:T356291 Reliable Trove backups
- phab:T353356 Magnum PoC with RelEng
- phab:T347490 [wmcs-cookbooks] Downtime alerts from cloudcumins
Data Services
- phab:T352206 [toolsdb] Upgrade to MariaDB 10.6
- phab:T344717 [toolsdb] test creating a new replica host
- phab:T344719 [toolsdb] test failover procedure

Q1/Q2 (Sep 2023 to Dec 2023)[edit]

Phabricator board

Availability[edit]

Komla out all Quarter
Arturo out all of August
David out for half of July, one week of August

Goals[edit]

Toolforge
- phab:search/query/WVN9f5IZgGkm/ Continue work on Envvars service (David) Done
- phab:T335249 Toolforge build service beta round 2 (David) Done
- phab:T334081 NFS-free webservice (David/Taavi) Done
- phab:T267374 Migration of projects from grid (Mostly communication, not active hand-holding this quarter) Done
- phab:T341084 Continuous delivery (gitlab pipeline things) (Arturo/David) Done
- Organize toolforge-wide workgroup to replace subtask-specific workgroups. (David) Done
- Reduce and define the UI surface of the toolforge platform via CLI and API definitions. (David + Taavi + Arturo)
- phab:T298005 Kubernetes version updates (Taavi + Arturo) Done
Superset
- phab:tag/superset.wmcloud.org/ Continue operational exploration and responding to feedback (Rook) Done
Cloud-vps + infra
- phab:T341285 Openstack Upgrade to A (Francesco + Andrew) Done
- phab:tag/openstack-magnum/ Magnum
  - phab:T328711 Experiment with magnum-ui, decide w/not it’s useful in current state (Rook + Andrew) Done
  - Automated testing (Rook and/or Andrew)
  - phab:T328712 Prepare service for non-wmcs users (Rook + Andrew)
- Trove
  - Automated testing (Rook and/or Andrew)
  - phab:T337396 Investigate possible postgres improvements (Francesco)
  - Continue to shepherd upstream fixes (Andrew)
    - https://review.opendev.org/c/openstack/trove/+/875262?usp=dashboard
    - https://review.opendev.org/c/openstack/trove/+/869511?usp=dashboard
- NFS
  - phab:T292546 Continue to improve/stabilize/expand cinder backups (Andrew) Done
  - phab:T333477 Review/improve observability (Andrew + Taavi) Done
- Network redesign
  - phab:T338937 Migrate swift to new network setup in codfw1dev (Arturo and Andrew) Done
  - phab:T341060 Design Eqiad implementation (Arturo)
- Ceph
  - phab:T309789 Host OS upgrades (Francesco + Raymond)
  - phab:T306820 Ceph Version upgrade (Post Host OS Upgrade) (David + Raymond)
  - phab:T297083 Multirack HA (David)
Team infra
- Cumin
  - phab:T319401,phab:T325067 Finish work on dedicated cumin exec hosts (Francesco) Done
- Observability
  - Continue to improve VM+prometheus+alert manager integration (Taavi)
  - phab:T328502 Move more alerts from icinga to alert manager (Taavi)
  - Enable silences from alerts.w.o for metricsinfra alerts (depends on o11y team upgrades to alertmanager)
Hackathon Support (Slavina) Done
phab:T327319 Documentation (Tricia, support from Francesco) Done