Deadlock detection and prevention in PostgreSQL - Time & Space Complexity

Time Complexity: Deadlock detection and prevention

O(n + e)

Understanding Time Complexity

When databases handle many tasks at once, sometimes they get stuck waiting for each other. This is called a deadlock.

We want to understand how the system finds and stops these deadlocks as more tasks run.

Scenario Under Consideration

Analyze the time complexity of PostgreSQL's deadlock detection query.

WITH RECURSIVE waiting_graph AS (
  SELECT
    blocked.pid AS blocked_pid,
    blocking.pid AS blocking_pid
  FROM pg_locks blocked
  JOIN pg_locks blocking ON blocked.locktype = blocking.locktype
    AND blocked.database IS NOT DISTINCT FROM blocking.database
    AND blocked.relation IS NOT DISTINCT FROM blocking.relation
    AND blocked.page IS NOT DISTINCT FROM blocking.page
    AND blocked.tuple IS NOT DISTINCT FROM blocking.tuple
    AND blocked.virtualxid IS NOT DISTINCT FROM blocking.virtualxid
    AND blocked.transactionid IS NOT DISTINCT FROM blocking.transactionid
    AND blocked.classid IS NOT DISTINCT FROM blocking.classid
    AND blocked.objid IS NOT DISTINCT FROM blocking.objid
    AND blocked.objsubid IS NOT DISTINCT FROM blocking.objsubid
  WHERE blocked.granted = false AND blocking.granted = true
),
search_cycle AS (
  SELECT blocked_pid, blocking_pid, ARRAY[blocked_pid] AS path
  FROM waiting_graph
  UNION ALL
  SELECT wg.blocked_pid, wg.blocking_pid, path || wg.blocked_pid
  FROM waiting_graph wg
  JOIN search_cycle sc ON wg.blocked_pid = sc.blocking_pid
  WHERE NOT wg.blocked_pid = ANY(path)
)
SELECT * FROM search_cycle WHERE blocking_pid = ANY(path[1:array_length(path,1)-1]);
    

This query looks for cycles in the waiting graph to detect deadlocks among transactions.

Identify Repeating Operations

Look for repeated checks and traversals in the query.

• Primary operation: Recursive search through the waiting graph to find cycles.
• How many times: The recursion repeats for each edge in the graph until a cycle is found or all paths are checked.

How Execution Grows With Input

The more transactions and locks there are, the bigger the waiting graph becomes.

Input Size (n)	Approx. Operations
10	About 20-30 recursive checks
100	Hundreds of recursive checks
1000	Thousands to tens of thousands of checks

Pattern observation: The work grows quickly as the number of waiting relationships grows, because the system must explore many paths to find cycles.

Final Time Complexity

Time Complexity: O(n + e)

This means the detection time grows roughly with the number of transactions and their waiting connections.

Common Mistake

[X] Wrong: "Deadlock detection only checks a fixed number of transactions, so it always runs fast."

[OK] Correct: The detection must explore all waiting links between transactions, which can grow a lot as more transactions wait on each other.

Interview Connect

Understanding how deadlock detection scales helps you explain how databases keep running smoothly even with many users.

Self-Check

"What if the database used a simpler method that only checked direct waits, not recursive cycles? How would the time complexity change?"