SHORT_POLLING_FETCH_STATUS.md

c3vis Request Handling & Caching

Short Polling for instance_summaries_with_tasks API

The following diagram depicts the interaction between the client and server
when retrieving cluster data.

When a client first asks the server for instances and tasks for a cluster (or on the first request after cluster state has expired), the server will:

1. Create a ClusterState object with FetchStatus = INITIAL
2. Store it in the in-memory ClusterStateCache
3. Return INITIAL to client
4. Proceed to process the request in the background ("Background Processing" swimlane)
5. Background: Populate the cached ClusterState object with cluster data
6. Background: Update the ClusterState object's FetchStatus
7. Return current state to client

Client will poll the server until it finds INITIAL or ERROR state.

Force Refresh

If the user clicks "Refresh Server Cache", the client will send a forceRefresh=true parameter on its first polling attempt. The server upon seeing forceRefresh=true will invalidate the cached ClusterState entry and begin the above process again.

Fetch Status State Diagram

Fetch Status State	Description
INITIAL	State object created
FETCHING	Server has begun fetching cluster data asynchronously
FETCHED	Server has completed fetching cluster data
ERROR	Server encountered an error fetching cluster data

Cluster State Cache Expiry and Fetch Status

When a ClusterState object is added to the in-memory cache, it is set to expire within clusterStateCacheTtl milliseconds. The default being 30mins. After that time, the cache object is set to null and won't be set to INITIAL until necessitated by a subsequent client request.