Audit of services and differences between `main` and `canary` stage in all environments
As part of the pipeline reorder epic &608 (closed) the canary
stage of staging
has now become the first line of testing before code goes to production canary
stage. This issue is to track the results of an audit to both the main
and canary
stage across both staging
and production
environments, to make sure we have a concrete and up to date list of what services are being deployed in which environment, and how they work with regards to canarying.
Service | In main stage |
In canary stage |
Canary by % traffic | Able to force canary | Issues related to getting canary for service |
---|---|---|---|---|---|
gitlab-pages | N/A | ||||
gitlab-shell | N/A | ||||
nginx-ingress |
|
N/A | |||
registry |
|
N/A | |||
webservice-api |
|
N/A | |||
webservice-git |
|
N/A | |||
webservice-web |
|
N/A | |||
webservice-websockets |
|
N/A | |||
praefect |
|
|
N/A | ||
Gitaly |
|
|
N/A | ||
Gitlab Agent for Kubernetes (KAS) | #1547 (closed) | ||||
mailroom | #2326 | ||||
sidekiq-catchall | gitlab-org/gitlab#19239 | ||||
sidekiq-database-throttled | gitlab-org/gitlab#19239 | ||||
sidekiq-elasticsearch | gitlab-org/gitlab#19239 | ||||
sidekiq-gitaly-throttled | gitlab-org/gitlab#19239 | ||||
sidekiq-imports | gitlab-org/gitlab#19239 | ||||
sidekiq-low-urgency-cpu-bound | gitlab-org/gitlab#19239 | ||||
sidekiq-memory-bound | gitlab-org/gitlab#19239 | ||||
sidekiq-quarantine | gitlab-org/gitlab#19239 | ||||
sidekiq-urgent-cpu-bound | gitlab-org/gitlab#19239 | ||||
sidekiq-urgent-other | gitlab-org/gitlab#19239 |
Edited by Graeme Gillies