The library documents itself · Concurrency & distribution
stzClusterSupervisor
13 own methods · 0 forms written with extensions, listed under their method
the source: base/cluster/stzClusterSupervisor.ring(opens in a new tab) · All classes · The area reference: Concurrency & distribution
Under each method, its extensions (Q CS XT Z ZZ IB W U): bold and ticked, it exists; struck, it does not. The extensions and what they mean
No method of this class is run on this page: the class reaches files, the network, processes, the clock, chance or sound, and an example run here would only be an example of its effects. Each method carries the explanation the library gives of itself.
Method
Explanation, as the source carries it
init
(no comment in the source: the name reads as a sentence) initno extension
Name_
(no comment in the source: the name reads as a sentence) name_no extension
SetName
(no comment in the source: the name reads as a sentence) set nameno extension
ClusterQ
(no comment in the source: the name reads as a sentence) cluster qno extension
Policy
policy ------------------------------------------------------------- Bound a profile's fleet size: never below nMin, never above nMax.no extension
SetWaterMarks
(no comment in the source: the name reads as a sentence) set water marksno extension
ReportLoad
The app's observed demand for a profile (0..1: e.g. latency/SLA or in-flight/capacity). This is the honest load signal -- the app measures it; the supervisor enforces the policy.no extension
LoadOf
(no comment in the source: the name reads as a sentence) load ofno extension
FeedLoadFrom
The perf-P6 closing of the loop: the load signal MEASURED instead of hand-fed. A stzPerfProfile's LoadRatio() is X/Xmax -- how close the observed workload is to its CPU ceiling -- read from the same engine instruments everything else trusts. Call it before each Supervise()/Cycle() pass (or from the same host tick).no extension
Decide
the decision (PURE policy) ----------------------------------------- Given the current fleet metrics + reported load, return the actions to take: [ [ :restart, tag ], [ :scaleup, tag ], [ :scaledown, tag ] ]. Dead-worker restart takes precedence; then load-driven scaling within the [min, max] bounds. No side effects -- Supervise() applies these.no extension
Supervise
applying the decision to the live cluster -------------------------- One supervision pass: refresh health, decide, apply. Returns the actions taken.no extension
LastActions
(no comment in the source: the name reads as a sentence) last actionsno extension
Cycle
R5 composition: a supervised job on stzAgentHost ------------------- stzAgentHost ticks Cycle() on any supervised object. So the cluster supervisor runs on the SAME reactor loop the fleet runs on. Returns the number of actions taken this tick (0 -> quiescent).no extension