Hi everyone, I have pipelines stuck in `STARTING` ...
# ask-community
n
Hi everyone, I have pipelines stuck in
STARTING
mode. They keep on accumulating over the period of time. Currently I have to terminate them manually from DAGIT UI. I want to terminate them programmatically using the Graphql api. Is there a way this can be achieved? I am using
pipelineRunsOrError
query to filter out the ones that are in
STARTING
mode, but I can't tell how long they have been in that state. I would like to terminate the ones that pending for more than a day.
Also, there is an
updatedAfter
argument in the
pipelineRunsFilter
. What type of value can be passed to it? I am hoping something like datetime could solve my issue
d
Hi Navneet - this doesn't directly answer your question but in later version of dagster we can automatically do this for you - it's possible that upgrading will be a similar amount of work to building this feature yourself? (start_timeout_seconds here: https://docs.dagster.io/deployment/run-monitoring#run-monitoring)
to more directly answer your question - there should be a startTime field on the run object in graphql
n
But startTime field for the runs in STARTING mode has value null.
d
We show a timestamp in dagit though right? Dagit uses graphql too so if it shows up in dagit there must be some way to fetch it
One sec
n
Hi, did you find something?
d
looks like back in 0.12.x this was pipelineRunsOrError { stats { startTime } }
Would still recommend considering an upgrade
n
Thanks, launchTime parameter works for me.