Does QuestDB cache query results?

Are query results cached in QuestDB ?
What does i.q.c.p.PGConnectionContext query cache used [fd=34248] appearing in logs mean ?

When I'm running

SELECT colX FROM (SUBQUERY)

for the first time it takes 1 second. On the next query it takes only 50 milliseconds.

QuestDB doesn't cache query results .
Running a query is likely to fetch data from storage to OS page cache, warm up the data. On the second run it's possible that whole data set is in RAM and there's no need to wait on IO operations.

To speed up query parsing, analysis and query plan building, QuestDB does cache execution plans.
i.q.c.p.PGConnectionContext query cache used [fd=34248] means that compatible plan was found in pgwire cache and QuestDB can skip aforementioned query execution stages.

You can adjust query cache sizes with the following server.conf settings :

HTTP

# enables the query cache
#http.query.cache.enabled=true

# sets the number of blocks for the query cache. Cache capacity is number_of_blocks * number_of_rows
#http.query.cache.block.count=4

# sets the number of rows for the query cache. Cache capacity is number_of_blocks * number_of_rows
#http.query.cache.row.count=16

PG wire protocol

# enables select query cache
#pg.select.cache.enabled=true

# sets the number of blocks for the select query cache. Cache capacity is number_of_blocks * number_of_rows
#pg.select.cache.block.count=16

# sets the number of rows for the select query cache. Cache capacity is number_of_blocks * number_of_rows
#pg.select.cache.row.count=16

# enables insert query cache
#pg.insert.cache.enabled=true

# sets the number of blocks for the insert query cache. Cache capacity is number_of_blocks * number_of_rows
#pg.insert.cache.block.count=8

# sets the number of rows for the insert query cache. Cache capacity is number_of_blocks * number_of_rows
#pg.insert.cache.row.count=8

Yes, QuestDB uses a query cache to improve query performance. The i.q.c.p.PGConnectionContext query cache message appearing in the logs indicates that a query result was retrieved from the cache rather than being executed again.

The query cache works by storing the results of SELECT queries in memory, so that if the same query is executed again, the result can be retrieved from the cache rather than executing the query again. This can significantly improve query performance for frequently executed queries.

In your case, it appears that the first time you ran the query, QuestDB had to execute the query and retrieve the result, which took 1 second. However, on the second query, QuestDB was able to retrieve the result from the query cache, which only took 50 milliseconds.

Note that the query cache is only effective for SELECT queries that return the same result for the same input parameters. If the underlying data changes, or if the input parameters to the query change, then the query cache will not be used and the query will need to be executed again.