analyze-traces.sh

#!/bin/bash
set -ex
set -o pipefail
trap "exit" INT TERM
trap "kill $(jobs -pr) ||:" EXIT

stage=${stage:-}
script_dir="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"

rm -rf data {affected-queries,shm-traces,trace-filtered,flamegraph-query-runs}.tsv {interesting-addrs-,stacks.}{left,right}.{tsv,rep} ||:

for version in right
do
rm -rf data
clickhouse-local --stacktrace --verbose --query "

create table addresses_src($(cat "$version-addresses.tsv.columns"))
    engine File(TSVWithNamesAndTypes, '$version-addresses.tsv');

create table addresses_join engine Join(any, left, addr) as
    select addr, name from addresses_src;

create table interesting_addrs engine File(TSVWithNamesAndTypes, 'interesting-addrs-$version.tsv') as
    select distinct addr from addresses_src
    where multiSearchAny(name,
        ['arena_maybe_decay']
        --['madvise', 'mmap', 'pages_commit_impl', 'arena_maybe_decay']
        )
    ;

select count(*) from file('$version-trace-log.tsv', TSVWithNamesAndTypes, '$(cat "$version-trace-log.tsv.columns")')
    where query_id = ''
    ;

create table stacks engine File(TSV, 'stacks.$version.rep') as
    select
        arrayStringConcat(
            arrayMap(
                x -> joinGet(addresses_join, 'name', x),
                trace
            ),
            ';'
        ) readable_trace,
        count()
    from file('$version-trace-log.tsv', TSVWithNamesAndTypes,
        '$(cat "$version-trace-log.tsv.columns")') trace_log
    where arrayJoin(trace) in interesting_addrs
        and query_id != ''
    group by trace
    ;
"

~/fg/flamegraph.pl < "stacks.$version.rep" > "$version.svg"

done