You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Once we have Task Queue support implemented, we need to auto-populate the queue as new package versions are released.
However, we don't want to burn CPU trying to rebuild every package in every ecosystem. We need to define some criteria for which packages are worth attempting to rebuild. This set of packages can be described as "tracked".
Some potential criteria:
The package must be in the top N packages for that ecosystem, where we grow the number N over time
The package must have successfully rebuilt at least once before
Initially, we can run our benchmarks manually and then query for successes. This would generate a static list of packages which we can use when filtering incoming package publish events.
The text was updated successfully, but these errors were encountered:
Once we have Task Queue support implemented, we need to auto-populate the queue as new package versions are released.
However, we don't want to burn CPU trying to rebuild every package in every ecosystem. We need to define some criteria for which packages are worth attempting to rebuild. This set of packages can be described as "tracked".
Some potential criteria:
Initially, we can run our benchmarks manually and then query for successes. This would generate a static list of packages which we can use when filtering incoming package publish events.
The text was updated successfully, but these errors were encountered: