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
As #57 mentions, we need a way to persist data between processes. Ideally we want something we can append to so we're not dumping the entire file every time. The specific goal of this issue is to make it possible to facilitate storage of the above.
Some things we should consider or already intend on storing:
Important events such as:
Error messages
Shutdowns
Restarts
Agent Upgrades
Resource, such as ram and cpu, usage history
Results or changes made by scheduled events
Any other kind of data where logging does not provide granular enough information or where we want data to be available for debugging and metrics purposes.
The text was updated successfully, but these errors were encountered:
I think for just making sure tasks don't end up in limbo because of an agent restart, it's sufficient if the agent includes its current list of tasks in every announce and reannounce. The master can then easily catch a discrepancy between that list and what it has in its database.
No persistent storage on the agent side is needed for that.
Yeah you're probably right that it will be sufficient enough, I've remove that point from this issue. I'm also moving this issue off of 0.8.0 because this is an enhancement that we don't yet have an immediate need for.
As #57 mentions, we need a way to persist data between processes. Ideally we want something we can append to so we're not dumping the entire file every time. The specific goal of this issue is to make it possible to facilitate storage of the above.
Some things we should consider or already intend on storing:
The text was updated successfully, but these errors were encountered: