The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Screen Shot 2022-08-17 at 1.28.03 PM.png

Hi I'm having some problems with the `QueuedRunCoordinator`. In the doc, it says by default it can have 10 concurrent runs but I can't seem to get pass 2. I scheduled two runs for a test. The first started running in a few seconds but the second one just stayed queued until the first one finished execution (about 10 minutes). I wonder if I need to explicitly configure the concurrency limit or is it related to something else? Thank you!

Some info about other configurations:
• Dagster version: 0.14.20
• RunLauncher: DefaultRunLauncher
• RunCoordinator: QueuedRunCoordinator (using the default config)
• Code Repo server size: 2vCPU and 8 GB of memory (Should I get a larger machine?)
• Dagit and Deamon server size: same as above.

Update: once I increased the limit to 25, I could run more jobs in parallel. It turns out that if I go to the Run menu and select `in progress` tab, I found a few orphaned runs that were still in progress even though they were submitted months ago. My assumption is that they are disrupted during deployments or there are name changes so they are just stuck...I probably need to add some timeout to the job.

Thanks for the update. Glad that you were able to figure it out