@Marvin Does prefect record the last successful run of a flow and if so how would I reference that date and time?

Prefect does keep track of the details of flow runs, including start times, completion times, and whether they completed successfully. In the documentation, there's an example that shows how to use the Prefect client to get information about the most recent flow runs:

python
import asyncio
from prefect import get_client

async def get_most_recent_flow_runs():
    async with get_client() as client:
        return await client.read_flow_run(where={"state": {"_eq": "Completed"}})

last_3_flow_runs = asyncio.run(get_most_recent_flow_runs())
print(last_3_flow_runs)
assert all(run.state.is_completed() for run in last_3_flow_runs)
assert (end_times := [run.end_time for run in last_3_flow_runs]) == sorted(end_times, reverse=True)

This code snippet retrieves the last three flow runs and asserts that they are completed and sorts them by end time in reverse order to ensure the most recent runs are listed first. For more details and examples of using filters and querying run data, you can refer to the relevant guides on using the client and filters API reference: - Using the Client - Filters API Reference (not directly linked but mentioned in the excerpt) Remember, if you want to narrow down the runs of a specific flow or deployment, you can adjust the query passed to

client.read_flow_run

using appropriate filters such as

DeploymentFilter

.

@Marvin The where clause in the code snippet is throwing an error.

It looks like there are no current open issues on the Prefect GitHub repository related to errors with the

where

clause in the client's

read_flow_run

method. It's possible that the code snippet I provided may not be exactly appropriate for your version of Prefect or there may be a missing detail. If you're having trouble with the

where

clause, ensure you're using it in the context of an async function and the client from the correct version of Prefect. Remember that your API query structure needs to match the schema expected by Prefect's GraphQL API. To resolve the issue, you could either refer to the Prefect documentation on API queries or post a GitHub issue for the Prefect community if the problem persists. Here's the adjusted version of the code snippet using the Prefect Client, without assuming any

where

clause issues:

python
import asyncio
from prefect import Client

async def get_latest_successful_flow_run():
    async with Client() as client:
        # Adjust the filter below according to your needs
        flow_runs = await client.read_flow_run()
        # Filter or process the flow_run as needed to find the last successful run
        # e.g., finding the last successful run by checking the `state` attribute

    # Return or print information about the latest successful flow run
    return flow_runs

# If using this, need to run it within an async context
latest_successful_run = asyncio.run(get_latest_successful_flow_run())
print(latest_successful_run)

Please note that you need to ensure that the filtering for the "successful" runs matches the schema and data structure expected by Prefect. If there are details about the error you're experiencing, I could potentially give you more targeted assistance. For more information on the Prefect Client and handling flow runs, please refer to Prefect Docs - Using the Client.