When an entry is created, a series of slots are created as well. These slots have different responsibilities, some for tracing, some for collecting and calculating run-time information, some for flow control, some for circuit breaking, and so on. You can customize your own logic by adding slots based on the following fundamental slots.

image

NodeSelectorSlot: Trace nodes and form a calling tree in the memory

  1. ContextUtil.enter("entrance1", "appA");
  2. Entry nodeA = SphU.entry("nodeA");
  3. if (nodeA != null) {
  4.     nodeA.exit();
  5. }
  6. ContextUtil.exit();

The code above will generate the following structure in memory:

  1.           machine-root
  2.                  /     
  3.                 /
  4.          EntranceNode1
  5.               /
  6.              /   
  7.       DefaultNode(nodeA)

Each DefaultNode is identified by both the context and resource name. In other words, one resource id may have multiple DefaultNode distinguished by the entry names declared in ContextUtil.enter(contextName).

  1.   ContextUtil.enter("entrance1", "appA");
  2.   Entry nodeA = SphU.entry("nodeA");
  3.   if (nodeA != null) {
  4.     nodeA.exit();
  5.   }
  6.   ContextUtil.exit();
  7.   ContextUtil.enter("entrance2", "appA");
  8.   nodeA = SphU.entry("nodeA");
  9.   if (nodeA != null) {
  10.     nodeA.exit();
  11.   }
  12.   ContextUtil.exit();

The code above will generate the following structure in memory:

  1.                    machine-root
  2.                    /         \
  3.                   /           \
  4.           EntranceNode1   EntranceNode2
  5.                 /               \
  6.                /                 \
  7.        DefaultNode(nodeA)   DefaultNode(nodeA)

The calling trace can be displayed by calling curl http://localhost:8719/tree?type=root.

  1. EntranceNode: machine-root(t:0 pq:1 bq:0 tq:1 rt:0 prq:1 1mp:0 1mb:0 1mt:0)
  2. -EntranceNode1: Entrance1(t:0 pq:1 bq:0 tq:1 rt:0 prq:1 1mp:0 1mb:0 1mt:0)
  3. --nodeA(t:0 pq:1 bq:0 tq:1 rt:0 prq:1 1mp:0 1mb:0 1mt:0)
  4. -EntranceNode2: Entrance1(t:0 pq:1 bq:0 tq:1 rt:0 prq:1 1mp:0 1mb:0 1mt:0)
  5. --nodeA(t:0 pq:1 bq:0 tq:1 rt:0 prq:1 1mp:0 1mb:0 1mt:0)
  7. t:threadNum  pq:passQps  bq:blockedQps  tq:totalQps  rt:averageRt  prq: passRequestQps 1mp:1m-passed 1mb:1m-blocked 1mt:1m-total

ClusterNodeBuilderSlot: Build cluster node named by resource

This slot maintains resource runtime statistics (response time, QPS, thread, count, exception), and a list of origin callers’ statistics. The origin caller’s name is marked by origin in ContextUtil.enter(contextName, origin).

The information can be displayed by calling the following HTTP API: http://localhost:8719/origin?id=xxxx

  1. id: nodeA
  2. idx origin  threadNum passedQps blockedQps totalQps aRt   1m-passed 1m-blocked 1m-total 
  3. 1   caller1 0         0         0          0        0     0         0          0        
  4. 2   caller2 0         0         0          0        0     0         0          0

StatisticSlot: Collect runtime statistics

  • ClusterNode (ResourceNode): Total statistics of a resource;
  • Origin node: Statistics of a ClusterNode from different callers;
  • DefaultNode: The entry of the node, marked by context entry name and resource ID;
  • The total sum statistics of incoming entrances.

The StatisticNode is backed by sliding window data structure (LeapArray):

sliding-window