New to GraphQL and trying to use the Prefect Interactive API; this is my query but I am getting an error, can anyone help? I have a feeling there is something super simple I am missing: query:

query {
  flow_run_state {
    flow_run {
      id,
      parameters,
      created,
      flow(where:{name:"my-flow-name"}) {
        name,
        id
      },
    }
  }
}

Error:

{
  "graphQLErrors": [
    {
      "message": "Unknown argument \"where\" on field \"flow\" of type \"flow_run\".",
      "extensions": {
        "code": "GRAPHQL_VALIDATION_FAILED"
      }
    }
  ],
  "networkError": null,
  "message": "GraphQL error: Unknown argument \"where\" on field \"flow\" of type \"flow_run\"."
}

Hi @Zach - I think that error is because you're applying the

where

clause on a nested field. Something like this should work instead:

query {
  flow_run_state(where: {flow_run: {flow: {name: {_eq: "my-flow-name"}}}}) {
    flow_run {
      id,
      parameters,
      created,
      flow {
        name,
        id
      },
    }
  }
}

ahhh so all your

where

clauses need to be at the top of the query?

It depends on how the resolver is written but that's often the case!

@nicholas One more question, is there anything I can do if my query keeps timing out?

Yes definitely, can you describe what you're trying to get? Maybe I can help you craft a performant query for it

I am trying to look at how long a certain task took across all flows that finished successfully. This is what I have so far:

query {
  flow_run_state(where: 
    {
      state: {_eq: "Success"},
      created: {_gt: "2020-09-11T23:09:48.215241+00:00"}
      flow_run: {
        task_runs: {task: {name: {_eq: "task-name"}}}
        flow: {name: {_eq: "flow-name"}}
      }
    }
  ) {
    state,
    flow_run {
      id,
      created,
      flow {
        name,
      },
      task_runs {
        task {
          name
        }
        start_time,
        end_time
      }
    }
  }
}

Ah ok, I can definitely see why that's timing out. For you knowledge, whenever you have nested fields, like:

flow_run { flow { id } }

Hasura creates a

JOIN

between the two tables. Sometimes this is performant and expected, but other times it's not, due to the size of the tables and/or the shape of the

ORDER_BY

and

WHERE

clauses it generates.

That worked! Thank you!

Fantastic! 👍