[Feature flag] Enable CI - Runner separation by plan
Summary
This issue is to rollout the runner's separation for certain plans on production,
that is currently behind the ci_runner_separation_by_plan
feature flag.
For the macOS SaaS Runners LA - which is planned for %15.0 release - we need to have a way to restrict the new macOS runners only to chosen plans. In the future we may use this feature to prepare more distinguished CI/CD execution environments for all three platforms we plan to support (Linux, macOS, Windows) and bound to specified SaaS plans.
How it works
When enabled, this feature will allow to define for any of the runner objects a list of plans that it supports
When set, only projects belonging to namespaces subscribed to one of the listed plans will be able to use such runner. Other jobs will be ignored by the scheduling algorithm.
When not set, which is the default, this feature is omitted and the plan is not considered when scheduling jobs to the runner that asked for it.
Owners
- Team: GitLab Runner group
- Most appropriate slack channel to reach out to:
#g_runner
- Best individual to reach out to: @tmaczukin
- PM: @DarrenEastman
Stakeholders
N/A
Expectations
What are we expecting to happen?
The expectation is that this feature will not allow not applicable projects - even if added manually to the allowlist - from executing jobs on the macOS SaaS Runner Managers. At the same time, we expect that this feature will not affect any other jobs nor create any performance decreases.
When is the feature viable?
-
Enable the feature by setting
ci_runner_separation_by_plan
feature flag to true. -
For the runner that we want to restrict with this feature, we need to set the list of plans that it should support. This can be done by Running a script like that in the production Rails console:
Ci::Runner.find(runner_id).update!(allowed_plans: ['premium', 'ultimate'])
The
runner_id
must be the valid ID of the runner we want to restrict.allowed_plans
array should contain valid plan names that are available on GitLab.com.
What might happen if this goes wrong?
No data loss is expected. The feature is quite simple and we don't expect it to do any harm. However, if something was missed during development, the worst thing that should happen is that some jobs become stuck because applicable runner (or runners) will reject the jobs.
A second potential problem, that is hard to be estimated if will happen or not (but we hope will not happen at all) is performance degradation of the jobs scheduling algorithm. The feature adds few more database requests that will be executed for some of the runners that ask for a job which in scale of GitLab.com may of course behave differently than what we've seen on the development instances.
In case of any of these problems happening or in case of any other problems that seem to be related to this feature, the mitigation is to disable the ci_runner_separation_by_plan
feature flag. This will turn-off the feature entirely. For macOS SaaS Runners - which we plan to use to test this feature - we will have a second layer of access limitation, so disabling this feature should not cause any security problems.
No MR reverting is needed.
What can we monitor to detect problems with this?
Consider mentioning checks for 5xx errors or other anomalies like an increase in redirects (302 HTTP response status)
https://dashboards.gitlab.net/d/ci-runners-incident-database/ci-runners-incident-support-database - will be useful to detect if there is any DB impact of the feature.
https://dashboards.gitlab.net/d/ci-runners-incident-gitlab-application/ci-runners-incident-support-gitlab-application - will be useful to detect if there are any unexpected errors happening when scheduling the jobs.
Both dashboards in their second row contain the same set of general metrics showing performance of jobs scheduling. If something will go wrong, these should also show the anomalies.
As we're considering DB performance problems and as there can be bugs not found during development, the feature when enabled may cause increase of 500 errors for the POST /api/v4/jobs/register
API requests.
What can we check for monitoring production after rollouts?
Consider adding links to check for Sentry errors, Production logs for 5xx, 302s, etc.
Dashboards mentioned above. General Rails production logs.
Rollout Steps
Rollout on non-production environments
- Ensure that the feature MRs have been deployed to non-production environments.
-
/chatops run auto_deploy status https://gitlab.com/gitlab-org/gitlab/-/merge_requests/83780
-
-
Enable the feature globally on non-production environments. -
/chatops run feature set ci_runner_separation_by_plan true --staging
-
-
Verify that the feature works as expected. Posting the QA result in this issue is preferable. The best environment to validate the feature in is staging-canary as this is the first environment deployed to. Note you will need to make sure you are configured to use canary as outlined here when accessing the staging environment in order to make sure you are testing appropriately.
Specific rollout on production
- Ensure that the feature MRs have been deployed to both production and canary.
-
/chatops run auto_deploy status https://gitlab.com/gitlab-org/gitlab/-/merge_requests/83780
-
The rest of standard tests in this section are not applicable, as the feature doesn't work in neither project, group nor user context. It works in context of the runner requesting a new job to be scheduled.
Preparation before global rollout
-
Check if the feature flag change needs to be accompanied with a change management issue. Cross link the issue here if it does. -
Ensure that you or a representative in development can be available for at least 2 hours after feature flag updates in production. If a different developer will be covering, or an exception is needed, please inform the oncall SRE by using the @sre-oncall
Slack alias. -
Ensure that documentation has been updated (More info). -
Announce on the feature issue an estimated time this will be enabled on GitLab.com. -
Ensure that any breaking changes have been announced following the release post process to ensure GitLab customers are aware. -
Notify #support_gitlab-com
and your team channel (more guidance when this is necessary in the dev docs).
Global rollout on production
For visibility, all /chatops
commands that target production should be executed in the #production
slack channel and cross-posted (with the command results) to the responsible team's slack channel (#g_runnerE
).
-
Incrementally roll out the feature. - If the feature flag in code has an actor, perform actor-based rollout.
[ ]/chatops run feature set <feature-flag-name> <rollout-percentage> --actors
- If the feature flag in code does NOT have an actor, perform time-based rollout (random rollout).
-
/chatops run feature set ci_runner_separation_by_plan 50 --random
-
- Enable the feature globally on production environment.
-
/chatops run feature set ci_runner_separation_by_plan true
-
- If the feature flag in code has an actor, perform actor-based rollout.
-
Announce on the feature issue that the feature has been globally enabled. -
Wait for at least one day for the verification term.
(Optional) Release the feature with the feature flag
If you're still unsure whether the feature is deemed stable but want to release it in the current milestone, you can change the default state of the feature flag to be enabled. To do so, follow these steps:
-
Create a merge request with the following changes. Ask for review and merge it. -
Set the default_enabled
attribute in the feature flag definition totrue
. -
Create a changelog entry.
-
-
Ensure that the default-enabling MR has been included in the release package. If the merge request was deployed before the monthly release was tagged, the feature can be officially announced in a release blog post. -
/chatops run release check <merge-request-url> <milestone>
-
-
Close the feature issue to indicate the feature will be released in the current milestone. -
Set the next milestone to this rollout issue for scheduling the flag removal. -
(Optional) You can create a separate issue for scheduling the steps below to Release the feature. -
Set the title to "[Feature flag] Cleanup ci_runner_separation_by_plan
". -
Execute the /copy_metadata <this-rollout-issue-link>
quick action to copy the labels from this rollout issue. -
Link this rollout issue as a related issue. -
Close this rollout issue.
-
WARNING: This approach has the downside that it makes it difficult for us to clean up the flag. For example, on-premise users could disable the feature on their GitLab instance. But when you remove the flag at some point, they suddenly see the feature as enabled and they can't roll it back to the previous behavior. To avoid this potential breaking change, use this approach only for urgent matters.
Release the feature
After the feature has been deemed stable, the clean up should be done as soon as possible to permanently enable the feature and reduce complexity in the codebase.
You can either create a follow-up issue for Feature Flag Cleanup or use the checklist below in this same issue.
-
Create a merge request to remove ci_runner_separation_by_plan
feature flag. Ask for review and merge it.-
Remove all references to the feature flag from the codebase. -
Remove the YAML definitions for the feature from the repository. -
Create a changelog entry.
-
-
Ensure that the cleanup MR has been included in the release package. If the merge request was deployed before the monthly release was tagged, the feature can be officially announced in a release blog post. -
/chatops run release check <merge-request-url> <milestone>
-
-
Close the feature issue to indicate the feature will be released in the current milestone. -
Clean up the feature flag from all environments by running these chatops command in #production
channel:-
/chatops run feature delete ci_runner_separation_by_plan --staging
-
/chatops run feature delete ci_runner_separation_by_plan
-
-
Close this rollout issue.
Rollback Steps
-
This feature can be disabled by running the following Chatops command:
/chatops run feature set ci_runner_separation_by_plan false