Integration / Apache Pulsar Interview questions
How does AppDynamics correlate a transaction across multiple tiers?
When a request enters the originating tier, the App Agent generates a unique identifier for that transaction instance and carries it forward as the request calls into downstream tiers, so every tier's activity for that one request can later be reassembled into a single, cross-tier view.
sequenceDiagram
participant Client
participant TierA as Tier A (originating)
participant TierB as Tier B (continuing)
participant Controller
Client->>TierA: HTTP request
TierA->>TierA: Agent generates correlation ID
TierA->>TierB: Outbound call + singularityheader
TierB->>TierB: Agent reads header, joins same transaction
TierA-->>Controller: Reports segment data
TierB-->>Controller: Reports segment data
Controller->>Controller: Reconstructs full cross-tier flow
For HTTP calls this identifier travels in a custom header named singularityheader; for other protocols like JMS it rides in an equivalent extensible field, such as a message property. The Controller doesn't do the correlating itself in real time — each tier reports its own segment, tagged with the same identifier, and the Controller reconciles those segments into one flow map after the fact.
What does each tier report to the Controller during a correlated transaction?
Which tier generates the initial correlation identifier?
More Related questions...
