The Autopsy: Profiling the Legacy Monolith

Every migration proposal starts with someone counting terabytes and tables. That count answers the wrong question. This section is the diagnostic work that has to happen before a single line of DDL gets translated: reading the legacy platform’s own instrumentation – Oracle AWR reports, Teradata DBQL, SQL Server Query Store – to find out which objects are actually driving load and cost, mapping stored-procedure dependencies and table access patterns so you know what breaks if you touch what, and rolling all of it into one workload inventory that becomes the source of truth every later decision – the 3-R call, the sequencing plan, the go/no-go matrix – traces back to. Skipping this autopsy is the single most common reason migrations blow their budget, so treat it as the foundation the rest of this part stands on, not a formality before the “real” work starts.

flowchart TD
    A["Legacy Monolith"] --> B["Instrumentation"]
    B --> B1["Oracle AWR"]
    B --> B2["Teradata DBQL"]
    B --> B3["SQL Server Query Store"]
    B1 & B2 & B3 --> C["Real bottlenecks:\ntop SQL, I/O waits, CPU hogs"]
    A --> D["Stored-procedure\ndependency graph"]
    A --> E["Table access\nheat map"]
    C & D & E --> F["Workload Inventory\n(source of truth)"]
    F --> G["3-R Decision\n(next section)"]

Lectures

#	Lecture	Est. Read
1	The Iceberg Assessment: 20% Data, 80% Everything Else	3 min read
2	Reading Oracle AWR Reports to Find the Real Bottlenecks	3 min read
3	Teradata DBQL and SQL Server Query Store: Same Job, Different Dialect	3 min read
4	Stored-Procedure Dependency Graphs and Table Heat Maps	4 min read
5	The Workload Inventory: Your Source of Truth	4 min read
6	Check Your Knowledge	Quiz

| ← Previous: Check Your Knowledge | Next: The Iceberg Assessment: 20% Data, 80% Everything Else → | |:—|—:|

The Autopsy: Profiling the Legacy Monolith

Lectures

Table of contents