Node Auto Scaling¶
Node auto scaling can be enabled by setting docker:auto-scale:enabled to true. It will try to add, remove and rebalance docker nodes used by tsuru.
Node scaling algorithms run in clusters of docker nodes, each cluster is based on the pool the node belongs to.
There are two different scaling algorithms that will be used, depending on how tsuru is configured: count based scaling, and memory based scaling.
Count based scaling¶
It’s chosen if docker:auto-scale:max-container-count is set to a value > 0 in your tsuru configuration.
Having max-container-count value as \(max\), the number of nodes in cluster as \(nodes\), and the total number of containers in all cluster’s nodes as \(total\), we get the number of free slots \(free\) with:
If \(free < 0\) then a new node will be added and tsuru will rebalance containers using the new node.
Having docker:auto-scale:scale-down-ratio value \(ratio\). tsuru will try to remove an existing node if:
Before removing a node tsuru will move it’s containers to other nodes available in the cluster.
To avoid entering loops, removing and adding node, tsuru will require \(ratio > 1\), if this is not true scaling will not run.
Memory based scaling¶
It’s chosen if docker:auto-scale:max-container-count is not set and your scheduler is configured to use node’s memory information, by setting docker:scheduler:total-memory-metadata and docker:scheduler:max-used-memory.
Having the amount of memory necessary by the plan with the largest memory requirement as \(maxPlanMemory\). A new node will be added if for all nodes the amount of unreserved memory (\(unreserved\)) satisfies:
Considering the amount of memory necessary by the plan with the largest memory requirement as \(maxPlanMemory\) and docker:auto-scale:scale-down-ratio value as \(ratio\). A node will be removed if its current containers can be distributed across other nodes in the same pool and at least one node still has unreserved memory (\(unreserved\)) satisfying:
Rebalancing containers will be triggered when a new node is added or if rebalancing would decrease the difference of containers in nodes by a number greater than 2, regardless the scaling algorithm.
Also, rebalancing will not run if docker:auto-scale:prevent-rebalance is set to true.
Auto scale events¶
Each time tsuru tries to run an auto scale action (add, remove, or rebalance). It will create an auto scale event. This event will record the result of the auto scale action and possible errors that occurred during its execution.
You can list auto scale events with tsuru-admin docker-autoscale-list
Running auto scale once¶
Even if you have docker:auto-scale:enabled set to false, you can make tsuru trigger the execution of the auto scale algorithm by running tsuru-admin docker- autoscale-run.