squash

squash is the web frontend to embed the bokeh apps and navigate through them. You can learn more about SQuaSH at SQR-009.

Requirements

The squash web frontend requires the squash-restful-api and squash-bokeh microservices, and the TLS certificats that are installed by the squash-deployment.

Kubernetes deployment

You can provision a Kubernetes cluster in GKE, clone this repo and deploy the squash microservice using:

cd squash
TAG=latest make service deployment

The variables SQUASH_MONITOR_APP and SQUASH_BOKEH_APPS can be used to control the bokeh apps allowed in this deployment, for example:

export SQUASH_MONITOR_APP=code_changes
export SQUASH_BOKEH_APPS="code_changes AMx PAx"

their default values correspond to the production deployment.

Debugging

Use the kubectl logs command to view the logs of the nginx and dash containers:

kubectl logs deployment/squash-api nginx
kubectl logs deployment/squash-api dash

Use the kubectl exec to run an interactive shell inside a container. Use tab completion or kubectl get pods command to find the pod's name:

kubectl exec -it <squash pod> -c dash /bin/bash

Rolling out updates

Check the update history with:

kubectl rollout history deployment squash

Modify the squash image and then apply the new configuration for the kubernetes deployment:

# we need to setup the env for django to collect static files
virtualenv env -p python3
source env/bin/activate
pip install -r requirements.txt
TAG=latest make build push update

Check the deployment changes:

kubectl describe deployments squash

Scaling up the squash microservice

Use the kubectl get replicasets command to view the current set of replicas, and then the kubectl scale command to scale up the squash deployment:

kubectl scale deployments squash --replicas=3

or change the kubernetes/deployment.yaml configuration file and apply the new configuration:

kubectl apply -f kubernetes/deployment.yaml

Check the deployment changes:

kubectl describe deployments squash
kubectl get pods
kubectl get replicasets

Development workflow

You can install the dependencies for developing

Install the software dependencies

git clone  https://github.com/lsst-sqre/squash.git

cd squash

virtualenv env -p python3
source env/bin/activate
pip install -r requirements.txt

Run squash

export SQUASH_DASH_DEBUG=True
export SQUASH_BOKEH_URL=<suqash-bokeh url> # e.g. the one from squash-bokeh deployment
export SQUASH_API_URL=<squash-restful-api url> # e.g. the one from squash-restful-api deployment

The variables SQUASH_MONITOR_APP and SQUASH_BOKEH_APPS can be used to control the bokeh apps allowed in this deployment, for example:

export SQUASH_MONITOR_APP=code_changes
export SQUASH_BOKEH_APPS="code_changes AMx PAx"

python manage.py runserver

The squash will run at http://localhost:8000.

lsst-sqre / squash