Skip to content

G
gitlab-ce
Commit c3f9ba16 authored by Sofia Vistas's avatar Sofia Vistas Committed by svistas
Browse files
Options

Add run and debug instructions for monitor specs 

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)

Apply 1 suggestion(s) to 1 file(s)
parent 27a9cc6d
Hide whitespace changes
Inline Side-by-side
Showing with 125 additions and 0 deletions
doc/development/testing_guide/end_to_end/running_tests_that_require_special_setup.md
View file @ c3f9ba16
......@@ -134,3 +134,128 @@ Once you have finished testing you can stop and remove the Docker containers:
docker stop gitlab-gitaly-ha praefect postgres gitaly3 gitaly2 gitaly1
docker rm gitlab-gitaly-ha praefect postgres gitaly3 gitaly2 gitaly1
```
## Guide to run and debug monitor tests
## How to set up
To run the Monitor tests locally, against the GDK, please follow the preparation steps below:
1. Complete the [Prerequisites](https://gitlab.com/gitlab-org/gitlab-development-kit/-/blob/master/doc/howto/auto_devops/index.md#prerequisites-for-gitlab-team-members-only), at least through step 5. Note that the monitor tests do not require permissions to work with GKE because they use [k3s as a Kubernetes cluster provider](https://github.com/rancher/k3s).
1. The test setup deploys the app in a Kubernetes cluster, using the Auto DevOps deployment strategy.
To enable Auto DevOps in GDK, follow the [associated setup](https://gitlab.com/gitlab-org/gitlab-development-kit/-/blob/master/doc/howto/auto_devops/index.md#setup) instructions. If you have problems, review the [troubleshooting guide](https://gitlab.com/gitlab-org/gitlab-development-kit/-/blob/master/doc/howto/auto_devops/tips_and_troubleshooting.md) or reach out to the `#gdk` channel in the internal GitLab Slack.
1. Do [secure your GitLab instance](https://gitlab.com/gitlab-org/gitlab-development-kit/-/blob/master/doc/howto/auto_devops/index.md#secure-your-gitlab-instance) since it is now publicly accessible on `https://[YOUR-PORT].qa-tunnel.gitlab.info`.
1. Install the Kubernetes command line tool known as `kubectl`. Use the [official installation instructions](https://kubernetes.io/docs/tasks/tools/install-kubectl/).
You might see NGINX issues when you run `gdk start` or `gdk restart`. In that case, run `sft login` to revalidate your credentials and regain access the QA Tunnel.
## How to run
Navigate to the folder in `/your-gdk/gitlab/qa` and issue the command:
```shell
QA_DEBUG=true CHROME_HEADLESS=false GITLAB_ADMIN_USERNAME=rootusername GITLAB_ADMIN_PASSWORD=rootpassword GITLAB_QA_ACCESS_TOKEN=your_token_here GITLAB_QA_ADMIN_ACCESS_TOKEN=your_token_here CLUSTER_API_URL=https://kubernetes.docker.internal:6443 bundle exec bin/qa Test::Instance::All https://[YOUR-PORT].qa-tunnel.gitlab.info/ -- qa/specs/features/browser_ui/8_monitor/all_monitor_core_features_spec.rb --tag kubernetes --tag orchestrated --tag requires_admin
```
The following includes more information on the command:
-`QA_DEBUG` - Set to `true` to verbosely log page object actions.
-`CHROME_HEADLESS` - When running locally, set to `false` to allow Chrome tests to be visible - watch your tests being run.
-`GITLAB_ADMIN_USERNAME` - Admin username to use when adding a license.
-`GITLAB_ADMIN_PASSWORD` - Admin password to use when adding a license.
-`GITLAB_QA_ACCESS_TOKEN` and `GITLAB_QA_ADMIN_ACCESS_TOKEN` - A valid personal access token with the `api` scope. This is used for API access during tests, and is used in the version that staging is currently running. The `ADMIN_ACCESS_TOKEN` is from a user with admin access. Used for API access as an admin during tests.
-`CLUSTER_API_URL` - Use the address `https://kubernetes.docker.internal:6443` . This address is used to enable the cluster to be network accessible while deploying using Auto DevOps.
-`https://[YOUR-PORT].qa-tunnel.gitlab.info/` - The address of your local GDK
-`qa/specs/features/browser_ui/8_monitor/all_monitor_core_features_spec.rb` - The path to the monitor core specs
-`--tag` - the meta-tags used to filter the specs correctly
At the moment of this writing, there are two specs which run monitor tests:
-`qa/specs/features/browser_ui/8_monitor/all_monitor_core_features_spec.rb` - has the specs of features in GitLab Core
-`qa/specs/features/ee/browser_ui/8_monitor/all_monitor_features_spec.rb` - has the specs of features for paid GitLab (Enterprise Edition)
## How to debug
The monitor tests follow this setup flow:
1. Creates a k3s cluster on your local machine.
1. Creates a project that has Auto DevOps enabled and uses an Express template (NodeJS) for the app to be deployed.
1. Associates the created cluster to the project and installs GitLab Runner, Prometheus and Ingress which are the needed components for a successful deployment.
1. Creates a CI pipeline with 2 jobs (`build` and `production`) to deploy the app on the Kubernetes cluster.
1. Goes to Operation > Metrics menu to verify data is being received and the app is being monitored successfully.
The test requires a number of components. The setup requires time to collect the metrics of a real deployment.
The complexity of the setup may lead to problems unrelated to the app. The following sections include common strategies to debug possible issues.
### Deployment with Auto DevOps
When debugging issues in the CI or locally in the CLI, open the Kubernetes job in the pipeline.
In the job log window, click on the top right icon labeled as *"Show complete raw"* to reveal raw job logs.
You can now search through the logs for *Job log*, which matches delimited sections like this one:
```shell
------- Job log: -------
```
A Job log is a subsection within these logs, related to app deployment. We use two jobs: `build` and `production`.
You can find the root causes of deployment failures in these logs, which can compromise the entire test.
If a `build` job fails, the `production` job won't run, and the test fails.
The long test setup does not take screenshots of failures, which is a known [issue](https://gitlab.com/gitlab-org/quality/team-tasks/-/issues/270).
However, if the spec fails (after a successful deployment) then you should be able to find screenshots which display the feature failure.
To access them in CI, go to the main job log window, look on the left side panel's Job artifacts section, and click Browse.
### Common issues
**Container Registry**
When enabling Auto DevOps in the GDK, you may see issues with the Container Registry, which stores
images of the app to be deployed.
You can access the Registry is available by opening an existing project. On the left hand menu,
select `Packages & Registries > Container Registries`. If the Registry is available, this page should load normally.
Also, the Registry should be running in Docker:
```shell
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
f035f339506c registry.gitlab.com/gitlab-org/build/cng/gitlab-container-registry:v2.9.1-gitlab "/bin/sh -c 'exec /b…" 3 hours ago Up 3 hours 0.0.0.0:5000->5000/tcp jovial_proskuriakova
```
The `gdk status` command shows if the registry is running:
```shell
run: ./services/registry: (pid 2662) 10875s, normally down; run: log: (pid 65148) 177993s
run: ./services/tunnel_gitlab: (pid 2650) 10875s, normally down; run: log: (pid 65154) 177993s
run: ./services/tunnel_registry: (pid 2651) 10875s, normally down; run: log: (pid 65155) 177993s
```
Also, restarting Docker and then, on the Terminal, issue the command `docker login https://[YOUR-REGISTRY-PORT].qa-tunnel.gitlab.info:443` and use the GDK credentials to login. Note that the Registry port and GDK port are not the same. When configuring Auto DevOps in GDK, the `gdk reconfigure` command outputs the port of the Registry:
```shell
*********************************************
Tunnel URLs
GitLab: https://[PORT].qa-tunnel.gitlab.info
Registry: https://[PORT].qa-tunnel.gitlab.info
*********************************************
```
These Tunnel URLs are used by the QA SSH Tunnel generated when enabling Auto DevOps on the GDK.
**Pod Eviction**
Pod eviction happens when a node in a Kubernetes cluster is running out of memory or disk. After many local deployments this issue can happen. The UI shows that installing Prometheus, GitLab Runner and Ingress failed. How to be sure it is an Eviction? While the test is running, open another Terminal window and debug the current Kubernetes cluster by `kubectl get pods --all-namespaces`. If you observe that Pods have *Evicted status* such as the install-runner here:
```shell
$ kubectl get pods --all-namespaces
NAMESPACE NAME READY STATUS RESTARTS AGE
gitlab-managed-apps install-ingress 0/1 Pending 0 25s
gitlab-managed-apps install-prometheus 0/1 Pending 0 12s
gitlab-managed-apps install-runner 0/1 Evicted 0 75s
```
You can free some memory with either of the following commands: `docker prune system` or `docker prune volume`.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7