Hello Daniel,
We're using distributed locking and queueing capabilities of workflowcore, specifically the redis provider. Our intention has been to create a highly scalable distributed service which horizontally scales up as the number of workflows increase. Each workflow is very lightweight and spends about 95% of its time asynchronously awaiting network calls. Here are my questions:
Metrics for Horizontal Scaling
- We are consuming the Redis distributed lock and queueing provider. Would the number of items in either of these collections be good indicators to horizontally scale up?
- On rare occasions dangling items are left in Redis by workflow. I cannot confirm yet but I suspect they are redlock items. Dangling items in Redis could interfere with scaling, do you have any recommendations regarding this?
Hello Daniel,
We're using distributed locking and queueing capabilities of workflowcore, specifically the redis provider. Our intention has been to create a highly scalable distributed service which horizontally scales up as the number of workflows increase. Each workflow is very lightweight and spends about 95% of its time asynchronously awaiting network calls. Here are my questions:
Metrics for Horizontal Scaling