Though we can lazily initialize statistics spaces, there is still a
problem with read-only replicas. If the workload does not write to
the storage (only reads with SELECTs), then callbre is used and no
queries will be dispatched to the storage master. As a result, no
statistics spaces will be initialized on the storage failover group.

To deal with it, we add a new API method to initialize statistics
on the instance startup.

The query statistics samples the execution time of the opentelemetry's
spans among 1% of all the running queries. Collected data is stored
in the temporary spaces on each instance independently and can be
analyzed by user (with the local SQL) or exported somewhere.
The statistics spaces would be created by the sbroad itself if they
do not exist yet.

Implement opentelemetry instrumentation in sbroad.
How to test:

docker run --name jaeger -d --rm -p6831:6831/udp -p6832:6832/udp \
-p16686:16686 -p14268:14268 jaegertracing/all-in-one:latest

Then run a stress test with sbroad. The results would be available
at http://localhost:16686/

Igor Kuznetsov authored 2 years ago and Denis Smirnov committed 2 years ago

Now we compare query explain plans instead of to compare generate backend sql queries.

Alexey Protsenko authored 2 years ago and Igor Kuznetsov committed 2 years ago

feat: update gcc to 4.9, fix version for rust and cartridge cli, disable tarantool install from vk repo

Reduce the amoutn of allocations, make logic more clear.

Previously we had an error in the parameter binding order when
a query had sub-queries in filters or join conditions.

The main goal of this refactoring is to get a syntax data node
vector in the order for building SQL query in a more clear manner.

The internal bind_params() destructs parameter vector and reuses
its allocated values to bind them inside the IR tree. It is logical
to consume the parameters as they would be destroyed and we can't
reuse them any more.