Skip to content

stolostron/search-v2-api

Repository files navigation

search-v2-api

For development

  1. Clone the repo and open the folder cd search-v2-api
  2. Run make setup to set up certificates make setup
  3. Run make run make run
  4. Open the graphql link in browser.
  5. Issue a search query in the input. Example:
query{
    search(input: [
        {
            filters: [
                {
                    property: "kind",
                    values:[ "pod" ]
                }
            ],
        }
    ]) {
    		related{
          kind
          count
        }
        count 
    }
}

And the output appears on the right hand side.

Run within cluster

  1. Build the search-v2-api code and push to your private repo in quay.

     make docker-build

     docker tag search-v2-api:latest quay.io/<your_id>/search-v2-api:latest

     docker push quay.io/<your_id>/search-v2-api:latest

  2. Setup Postgres on your cluster using these instructions. Create postgres service by applying the yaml below.

apiVersion: v1
kind: Service
metadata:
  labels:
    postgres-operator.crunchydata.com/cluster: hippo
    postgres-operator.crunchydata.com/role: primary
  name: hippo-primary-ocm
  namespace: postgres-operator
spec:
  ports:
  - name: postgres
    port: 5432
    protocol: TCP
    targetPort: postgres
  selector:
    postgres-operator.crunchydata.com/cluster: hippo
    postgres-operator.crunchydata.com/role: master
  1. Replace the search-v2-api image deployment image with the image you pushed in Step 1. Note that you have to add your quay image pull secret to the deployment so that this private image can be pulled successfully.Update the TLS secret path in volumemount.

  2. Add the environment variables required for connection to Postgres database in the search-v2-api deployment.

      containers:
      - env:
        - name: DB_PASSWORD
          value: **************
        - name: DB_HOST
          value: hippo-primary-ocm.postgres-operator.svc.cluster.local
        - name: DB_USER
          value: hippo
        - name: DB_NAME
          value: hippo
        - name: DB_PORT
          value: "5432"
        image: quay.io/<your_id>/search-v2-api:latest
        name: search-v2-api
      imagePullSecrets:
      - name: my-quay-image-pull-secret
      - name: multiclusterhub-operator-pull-secret

Metrics

Search-v2-api also monitors and exports various metrics to Prometheus. Below are the metrics currently exported from search-v2-api.

Histograms:

  • search_api_requests - Histogram of HTTP requests duration (seconds).

    • Labels:
      • code: Status code generated from request
      • TODO: query_type: Type of query requested
  • search_dbquery_duration_seconds - Latency of DB requests in seconds.

    • Labels:
      • TODO: query_name: Name of database query

Example: The following metric will record all search queries created in under or equal to 0.1 seconds

search_api_requests_bucket{
    query_type="searchComplete",
    code="200",
    le="0.1"
}

Counters:

  • search_api_db_connection_failed_total - The total number of DB connection that has failed
    • Labels:
      • TODO:route: Route associated with DB connection failure.

To view these metrics, with the search api pod and database running, run the following command:

curl https://localhost:4010/metrics -k | grep search_


Rebuild Date: 2023-09-21

About

search api using postgres as store

Resources

Stars

Watchers

Forks

Packages

No packages published